Skip to main content

Full text of "Occupant Behavior, Egress, and Emergency Communication. Federal Building and Fire Safety Investigation of the World Trade Center Disaster (NIST NCSTAR 1-7)"

See other formats


NIST NCSTAR 1-7 

Federal Building and Fire Safety Investigation of the 
World Trade Center Disaster 

Occupant Behavior, Egress, and 
Emergency Communications 



Jason D. Averill 
Dennis S. Mileti 
Richard D. Peacock 
Erica D. Kuligowski 
Norman Groner 
Guylene Proulx 
Paul A. Reneke 
Harold E. Nelson 



nist 



National Institute of Standards and Technology • Technology Administration • U.S. Deportment of Commerce 



NIST NCSTAR 1-7 

Federal Building and Fire Safety Investigation of the 
World Trade Center Disaster 

Occupant Behavior, Egress, and 
Emergency Communications 



Jason D. Averill Norman Groner 

Building and Fire Research Laboratory John Jay College 

National Institute of Standards and Technology 

Guylene Proulx 
Dennis S. Mileti National Research Council Canada 

University of Colorado - Boulder 

Paul A. Reneke 
Richard D. Peacock Building and Fire Research Laboratory 

Erica D. Kuligowski National Institute of Standards and Technology 

Building and Fire Research Laboratory 
National Institute of Standards and Technology Harold E. Nelson 

Consultant 



September 2005 









^ , 



**TES O* ** 



U.S. Department of Commerce 
Carlos M. Gutierrez, Secretary 

Technology Administration 

Michelle O'Neill, Acting Under Secretary for Technology 

National Institute of Standards and Technology 
William Jeffrey, Director 



Disclaimer No. 1 

Certain commercial entities, equipment, products, or materials are identified in this document in order to describe a 
procedure or concept adequately or to trace the history of the procedures and practices used. Such identification is 
not intended to imply recommendation, endorsement, or implication that the entities, products, materials, or 
equipment are necessarily the best available for the purpose. Nor does such identification imply a finding of fault or 
negligence by the National Institute of Standards and Technology. 

Disclaimer No. 2 

The policy of NIST is to use the International System of Units (metric units) in all publications. In this document, 
however, units are presented in metric units or the inch-pound system, whichever is prevalent in the discipline. 

Disclaimer No. 3 

Pursuant to section 7 of the National Construction Safety Team Act, the NIST Director has determined that certain 
evidence received by NIST in the course of this Investigation is "voluntarily provided safety-related information" that is 
"not directly related to the building failure being investigated" and that "disclosure of that information would inhibit the 
voluntary provision of that type of information" (15 USC 7306c). 

In addition, a substantial portion of the evidence collected by NIST in the course of the Investigation has been 
provided to NIST under nondisclosure agreements. 

Disclaimer No. 4 

NIST takes no position as to whether the design or construction of a WTC building was compliant with any code 
since, due to the destruction of the WTC buildings, NIST could not verify the actual (or as-built) construction, the 
properties and condition of the materials used, or changes to the original construction made over the life of the 
buildings. In addition, NIST could not verify the interpretations of codes used by applicable authorities in determining 
compliance when implementing building codes. Where an Investigation report states whether a system was 
designed or installed as required by a code provision, NIST has documentary or anecdotal evidence indicating 
whether the requirement was met, or NIST has independently conducted tests or analyses indicating whether the 
requirement was met. 

Use in Legal Proceedings 

No part of any report resulting from a NIST investigation into a structural failure or from an investigation under the 
National Construction Safety Team Act may be used in any suit or action for damages arising out of any matter 
mentioned in such report (15 USC 281a; as amended by P.L. 107-231). 



National Institute of Standards and Technology National Construction Safety Team Act Report 1-7 
Natl. Inst. Stand. Technol. Natl. Constr. Sfty. Tm. Act Rpt. 1-7, 298 pages (September 2005) 
CODEN: NSPUE2 



U.S. GOVERNMENT PRINTING OFFICE 
WASHINGTON: 2005 



For sale by the Superintendent of Documents, U.S. Government Printing Office 
Internet: bookstore.gpo.gov — Phone: (202) 512-1800 — Fax: (202) 512-2250 
Mail: Stop SSOP, Washington, DC 20402-0001 



Abstract 



This report describes the occupant evacuation of World Trade Center (WTC) 1 and WTC 2 on 
September 11, 2001. Multiple sources of information were collected and analyzed: over 1,000 new 
interviews with survivors (including 803 telephone interviews, 225 face-to-face interviews, and 5 focus 
groups); over 700 published interviews; 9-1-1 emergency calls; transcripts of emergency 
communications, historical building design drawings, memoranda, and calculations; formal complaints 
filed with the Occupational Safety and Health Administration; and other relevant materials. 

The egress system, including stairwells and elevators, was described and compared to requirement of both 
contemporary and current code requirements. This report documents the emergency procedures, both as 
they were designed to be implemented, as well as how they were actually implemented on 
September 11,2001. 

The population in WTC 1 and WTC 2 on September 11, 2001, at 8:46:30 a.m. was enumerated and 
described, where the characteristics of the population were relevant to the subsequent evacuation, 
including training, experience, mobility status, among others. The progress of the evacuation of both 
towers was described in a quasi-chronological manner from 8:46:30 a.m. when WTC 1 was attacked, until 
10:28:22 a.m., when WTC 1 collapsed. 

Causal models were built to explore the sources of evacuation initiation delay (why people did not 
immediately start to leave the building) as well as normalized stairwell evacuation time (how long the 
average occupant spent in the stairwells per floor). Issues identified as contributing to either speeding or 
aiding the evacuation process were explored. Egress simulations provided context for estimating how 
long WTC 1 and WTC 2 would have taken to evacuate with different populations, using different models, 
and subject to different damage to the building. 

Keywords: Building fires, egress, egress modeling, emergency communication, evacuation, human 
behavior, interviews, World Trade Center. 



NIST NCSTAR 1-7, WTC Investigation iii 



Abstract 



This page intentionally left blank 



iv NIST NCSTAR 1-7, WTC Investigation 



Table of Contents 



Abstract iii 

List of Figures ix 

List of Tables xi 

List of Acronyms and Abbreviations xiii 

Metric Conversion Table xv 

Preface xix 

Acknowledgments xxix 

Executive Summary xxix 

Chapter 1 

Background and Introduction 1 

1.1 Historical Incidents in Which Egress Was Significant 1 

1.2 Previous Fires and Evacuation Incidents in the World Trade Center 4 

1.3 Scope of Occupant Behavior, Egress, and Emergency Communications Project for the 

NIST World Trade Center Investigation 6 

1.4 Report Organization 7 

Chapter 2 

Design of the World Trade Center Egress System 9 

2.1 Overall Building Description 9 

2.1.1 Description of the Towers 10 

2.2 Building Systems 18 

2.2.1 Egress Calculations 18 

2.2.2 Stairwells 27 

2.2.3 Elevators 32 

2.2.4 Emergency Communication System 35 

2.3 The Human Element 37 

2.3.1 Responsibilities of the Fire Safety Director 38 

2.3.2 Responsibilities of the Deputy Fire Safety Director 38 

2.3.3 Responsibilities of the Operations Control Center Supervisor 39 

2.3.4 Responsibilities of the Operations and Maintenance Management 39 

2.3.5 Floor Warden System 40 

2.3.6 Occupant and Tenant Training 40 



NISTNCSTAR 1-7, WTC Investigation 



Table of Contents 



2.4 Changes to the Egress System after the 1993 Bombing 41 

2.5 Building Code Analysis 43 

2.5.1 Egress in the Building Codes 43 

2.5.2 New York City Building Code (1968) 45 

2.5.3 New York City Building Code (October 2003) 46 

2.5.4 International Building Code (2000) 46 

2.5.5 NFPA 5000 and NFPA 101 - Life Safety Code (2003) 46 

2.5.6 Comparison of Current Code Requirements 47 

Chapter 3 

First-Person Data Collection and Analysis Methods 49 

3.1 Purpose and Scope of Collection of First-Person Data 49 

3.2 Methods 49 

3.2.1 Published First-Person Accounts 49 

3.2.2 Telephone Surveys 50 

3.2.3 Face-to-Face Interviews 59 

3.2.4 Focus Groups 63 

3.2.5 Audio, Video, Photographic, and Records Collection 64 

3.3 Analysis of First-Person Data 64 

Chapter 4 

September 11, 2001, Before the Attacks 67 

4.1 Building Population on September 11, 2001 67 

4.2 Occupant Characteristics 68 

4.3 Previous Experience 70 

4.4 Occupant Preparedness 70 

Chapter 5 

September 11, 2001, 8:46:30 a.m. Flight 11 Crashes into WTC 1 73 

5.1 Introduction 73 

5.2 Initial Observations in WTC 1 75 

Chapter 6 

September 11, 2001, 8:47 a.m. - 9:02 a.m. Occupants React to the Attack on WTC 1 79 

6.1 Conditions Worsen in WTC 1 79 

6.1.1 Activities and Information 81 

6.1.2 Emergency Response at the Fire Command Station, Lobby, WTC 1 85 



NIST NCSTAR 1-7, WTC Investigation 



Table of Contents 



6.1.3 Survivors Begin Their Evacuation 86 

6.2 Initial Observations and Reactions from WTC 2 89 

6.2.1 Activities and Information- WTC 2 91 

6.2.2 Survivors Begin Their Evacuation 95 

6.2.3 Elevator Use in WTC 2 96 

6.2.4 Announcements in WTC 2 98 

Chapter 7 

September 11, 2001, 9:02:59 a.m. Flight 175 Crashes into WTC 2 99 

Chapter 8 

September 11, 2001, 9:02:59 a.m. - 9:58:59 a.m. 56 Critical Minutes 105 

8.1 Survivors from at or above Floor 78 in WTC 2, After Impact 106 

8.2 Progress of Evacuation below the Impact Region 107 

8.3 Evacuation Route through Concourse 109 

8.4 More Announcements in WTC 2 110 

8.5 Conditions Continue to Deteriorate in Both Towers Ill 

Chapter 9 

September 11, 2001, 9:58:59 a.m. - 10:28:22 a.m. Collapse 115 

9.1 Evacuating WTC 1 115 

9.2 Location of Victims and Survivors within the Building 119 

9.2.1 Fatalities in Elevators 122 

Chapter 10 

Discussion and Analysis 125 

10.1 Overview of Total Evacuation Time 125 

10.1.1 Evacuation Initiation Delay Time in WTC 1 and WTC 2 127 

10.1.2 Time and Average Travel Speed in the Stairwells, WTC 1 129 

10.2 Factors Influencing Total Evacuation Time 130 

10.2.1 Predicting Evacuation Delay in World Trade Center 131 

10.2.2 Predicting Normalized Stairwell Evacuation Time in WTC 1 on 

September 11, 2001 135 

10.3 Issues That Impacted Occupant Evacuation 138 

10.3.1 Environmental Cues and Information 138 

10.3.2 Building Alarm Systems 141 

10.3.3 Public Address Announcements 142 

10.3.4 9-1-1 System 145 

NIST NCSTAR 1-7, WTC Investigation vii 



Table of Contents 



10.3.5 Previous Evacuation Experience 146 

10.3.6 Occupant Activities 147 

10.3.7 Aids and Constraints to Evacuation 151 

10.3.8 Emergency Responders and Building Authorities 153 

10.3.9 Occupant Experience 158 

10.3.10 Mobility-Impaired Occupants 158 

10.4 Evacuation Simulations 160 

10.4.1 Egress Simulation Results 160 

10.4.2 Egress Modeling Conclusions 164 

Chapter!! 

Findings 165 

11.1 Building Population and Demographics 165 

11.2 Evacuation 165 

11.3 Delays in Evacuation 167 

11.4 Emergency Management and Planning 168 

11.5 Occupants with Mobility Impairments 168 

Chapter 12 

References 171 

Appendix A 

Telephone Interview Script 175 

Appendix B 

Qualitative Analysis Coding 211 

Appendix C 

Causal Modeling 217 

Appendix D 

Egress Modeling 237 



viii NIST NCSTAR 1-7, WTC Investigation 



List of Figures 



Figure P-l. The eight projects in the federal building and fire safety investigation of the WTC 

disaster xxi 



Figure 2-1. WTC site plan 10 

Figure 2-2. Typical WTC tower architectural floor plan 11 

Figure 2-3. Shopping mall layout underneath WTC plaza 13 

Figure 2-4. WTC 1 lobby (concourse) level 14 

Figure 2-5. WTC 2 lobby (concourse) level 15 

Figure 2-6. Typical WTC tenant spaces 17 

Figure 2-7. A WTC 4 trading floor 17 

Figure 2-8. 1993 Windows on the World egress calculation, floor 106 25 

Figure 2-9. 1993 Windows on the World egress calculation, floor 107 26 

Figure 2-10. 44 in. stairwell in WTC 1 taken on September 11, 2001 27 

Figure 2-11. Stairwells in the WTC towers 29 

Figure 2-12. Horizontal transfer floors in the WTC towers 30 

Figure 2-13. Stairwell door signage as seen from inside stairwell A 30 

Figure 2-14. Elevator riser diagram for WTC 1 and WTC 2 33 

Figure 2-15. Fire command station in lobby of WTC on September 11, 2001, as seen from 

mezzanine escalator, looking west 37 

Figure 2-16. Elevator communication panel in the fire command station of WTC 1, as operated on 

September 11,2001 40 

Figure 4-1. Employment start date at WTC 69 

Figure 5-1. WTC 1 impact, 8:46 a.m 73 

Figure 5-2. Computer simulated impact damage to WTC 1 on floor 95 at 0.7 s after impact with 

stairwells superimposed 74 

Figure 5-3. Calculated damage to floors 93 through 98 in WTC 1 75 



Figure 6-1. Observations of building damage after initial awareness but before beginning 

evacuation in WTC 1 81 

Figure 6-2. Fire Command Station in WTC 1 on September 11,2001 86 



NIST NCSTAR 1-7, WTC Investigation ix 



List of Figures 



Figure 7-1. WTC 2 damage (computer simulated) at 0.62 s after impact with stairwells 

superimposed 99 

Figure 7-2. East-looking view through WTC 1 lobby as debris from WTC 2 impact travels past 

lobby windows onto plaza 100 

Figure 7-3. Calculated damage to floors 78 through 83 in WTC 2 (see NIST NCSTAR 1-6) 101 

Figure 7-4. Observations of damage from tenant spaces in WTC 2 103 

Figure 8-1. View from mezzanine level in WTC 1 looking east across WTC plaza, covered with 

debris. 'Sphere,' artwork by Fritz Koening, can be seen on the plaza 109 

Figure 8-2. Escalator from mezzanine to concourse level in WTC 1 on September 11, 2001 110 

Figure 8-3. Evacuees leaving complex near WTC 5 1 10 

Figure 8-4. Occupants seeking fresh air on floors 103 - 105 on north face of WTC 1 at 9:58:12 

a.m 112 

Figure 9-1. Collapse of WTC 2, viewed from north 116 

Figure 9-2. Occupants leaving a previously stuck elevator on September 11, 2001, in the lobby of 

WTC 1 123 

Figure 10-1. Reported evacuation time for survivors of WTC 1 and WTC 2 126 

Figure 10-2. Percentage of survivors remaining in the building for WTC 1 and WTC 2 126 

Figure 10-3. Percentage of occupants remaining in the building for WTC 1 and WTC 2 127 

Figure 10-4. Stairwell travel speed in WTC 1 for all stairwells 129 

Figure 10-5. Model of evacuation delay 132 

Figure 10-6. Causal model for predicting normalized stairwell evacuation time 136 

Figure 10-7. Firefighter and occupants using a 44 in. stairwell in WTC 1 on September 11, 2001 154 

Figure 10-8. Occupant and firefighter in bunker gear passing in a generic 44 in. stairwell 155 

Figure 10-9. Simulation of full building evacuation of a WTC tower with different occupant loads. ... 163 



NIST NCSTAR 1-7, WTC Investigation 



List of Tables 



Table P-l. Federal building and fire safety investigation of the WTC disaster xx 

Table P-2. Public meetings and briefings of the WTC Investigation xxiii 

Table 2-1. Use of floors in the WTC towers 12 

Table 2-2. Plan view of stairwells in WTC 1 and WTC 2 20 

Table 2-3. Minimum stairwell design for 42,850 ft 2 office plan 47 

Table 3-1. Comparison of media interviewees and badge list 51 

Table 3-2. Comparison of CNN victim list and badge list 52 

Table 3-3. Disposition of the CATI sample and the total sample by tower 54 

Table 3-4. Summary disposition rates by tower 55 

Table 3-5. Response rate analysis for WTC 1 59 

Table 3-6. Response rate analysis for WTC 2 59 

Table 3-7. Example tabular face-to-face interview data entry 61 

Table 4-1. Occupancy estimates on September 11, 2001, by tower 67 

Table 4-2. Occupancy estimate uncertainty 68 

Table 4-3. Age for telephone survey respondents in WTC 1 and WTC 2. a 68 

Table 4-4. WTC fire drills in 12 months prior to September 11, 2001. a 70 

Table 5-1. How survivors in WTC 1 became aware something was wrong on 

September 11,2001 76 

Table 5-2. Observations of building damage in WTC 1 when occupants first became aware 

something was wrong on September 11, 2001 78 

Table 6-1. Observations of conditions in WTC 1 before beginning evacuation 80 

Table 6-2. Activities prior to evacuation reported in telephone survey by survivors from WTC 1 82 

Table 6-3. Information received and sought prior to beginning evacuation in WTC 1 83 

Table 6-4. Survivors perception of risk to self and others after airplane impact and prior to 

entering stairwell in WTC 1 83 

Table 6-5. Sources of help used by occupants prior to beginning their evacuation in WTC 1 84 

Table 6-6. Single reason given by survivors to begin their evacuation in WTC 1 85 



NISTNCSTAR 1-7, WTC Investigation 



List of Tables 



Table 6-7. Elapsed time (min) to initiate evacuation for survivors from WTC 1 88 

Table 6-8. Stairwell chosen for evacuation in WTC 1 89 

Table 6-9. How survivors in WTC 2 became aware that something was wrong on September 11, 

2001 89 

Table 6-10. Activities prior to evacuation reported in telephone surveys by survivors from 

WTC 2 92 

Table 6-11. Information received and sought prior to beginning evacuation in WTC 2 93 

Table 6-12. Survivors' perception of risk upon airplane impact and prior to beginning their 

evacuation in WTC 2 93 

Table 6-13. How survivors received help prior to beginning their evacuation from WTC 2 94 

Table 6-14. Single reason given by survivors to begin their evacuation from WTC 2 94 

Table 6-15. Elapsed time (min) to initiate evacuation for survivors from WTC 2 95 

Table 6-16. Stairwell chosen for evacuation in WTC 2 96 

Table 8-1. Estimated times for survivors leaving the building in WTC 1 and WTC 2 105 

Table 8-2. Location of WTC 2 survivors at or above floors of impact at 9:03 a.m 106 

Table 8-3 . Observations of conditions during evacuation of WTC 1 and WTC 2 113 

Table 8-4. Total evacuation time (min) for survivors from WTC 2 114 

Table 9-1. Total evacuation time (min) for survivors from WTC 1 119 

Table 9-2. World Trade Center decedent location 120 

Table 10-1. Constraints to evacuation 151 

Table 10-2. Egress simulation matrix 161 

Table 10-3. Results for phased evacuation simulations 161 

Table 10-4. Total building evacuation time (simulated) for various occupant loads 162 

Table 10-5. Simulated evacuation results for fully-occupied WTC 1 and WTC 2 on 

September 11,2001 164 



xu 



NIST NCSTAR 1-7, WTC Investigation 



List of Acronyms and Abbreviations 



Acronyms 

ASTM ASTM International 

BPS Building Performance Study 

CATI Computer Assisted Telephone Interview 

ESU Emergency Services Unit 

FEMA Federal Emergency Management Agency 

FDNY City of New York Fire Department 

IBC International Building Code 

NFPA National Fire Protection Association 

NIST National Institute of Standards and Technology 

NYC New York City 

NYPD New York City Police Department 

OSHA Occupational Safety and Health Administration 

PANYNJ Port Authority of New York and New Jersey 

PAPD Port Authority Police Department 

WTC 1 World Trade Center 1 (North Tower) 

WTC 2 World Trade Center 2 (South Tower) 

WTC 7 World Trade Center 7 



Abbreviations 


°C 


degrees Celsius 


op 


degrees Fahrenheit 


ft 


foot 


ft 2 


square foot 


in. 


inch 


L 


liter 


m 


meter 


|am 


micrometer 


min 


minute 


s 


second 



NIST NCSTAR 1-7, WTC Investigation xiii 



List of Acronyms and Abbreviations 



This page intentionally left blank. 



NIST NCSTAR 1-7, WTC Investigation 



Metric Conversion Table 



To convert from 



to 



Multiply by 



AREA AND SECOND MOMENT OF AREA 



square foot (ft ) 
square inch (in. ) 
square inch (in. ) 
square yard (yd ) 



square meter (m ) 
square meter (m ) 
square centimeter (cm ) 
square meter (m ) 



9.290 304 E-02 
6.4516 E-04 
6.4516 E+00 
8.361 274 E-01 



ENERGY (includes WORK) 

kilowatt hour (kW ■ h) 

quad (1015 BtuIT) 

therm (U.S.) 

ton of TNT (energy equivalent) 

watt hour (W ■ h) 

watt second (W • s) 



joule (J) 
joule (J) 
joule (J) 
joule (J) 
joule (J) 
joule (J) 



3.6 E+06 
1.055 056 E+18 
1.054 804 E+08 
4.184 E+09 
3.6E+03 
1.0 E+00 



FORCE 

dyne (dyn) 
kilogram-force (kgf) 
kilopond (kilogram-force) (kp) 
kip (1 kip= 1,000 lbf) 
kip (1 kip= 1,000 lbf) 
pound-force (lbf) 



newton (N) 
newton (N) 
newton (N) 
newton (N) 
kilonewton (kN) 
newton (N) 



1.0E-05 
9.806 65 E+00 
9.806 65 E+00 
4.448 222 E+03 
4.448 222 E+00 
4.448 222 E+00 



FORCE DIVIDED BY LENGTH 

pound-force per foot (lbf/ft) 
pound-force per inch (lbf/in.) 



newton per meter (N/m) 
newton per meter (N/m) 



1.459 390 E+01 
1.751 268E+02 



HEAT FLOW RATE 

calorieth per minute (calth/min) 
calorieth per second (calth/s) 
kilocalorieth per minute (kcalth/min) 
kilocalorieth per second (kcalth/s) 



watt (W) 
watt (W) 
watt (W) 
watt (W) 



6.973 333 E-02 
4.184 E+00 
6.973 333 E+01 
4.184 E+03 



NISTNCSTAR 1-7, WTC Investigation 



Metric Conversion Table 



To convert from 



to 



Multiply by 



LENGTH 

foot (ft) 
inch (in) 
inch (in.) 
micron (m) 
yard (yd) 



meter (m) 
meter (m) 
centimeter (cm) 
meter (m) 
meter (m) 



3.048 E-01 
2.54 E-02 
2.54 E+00 
1.0E-06 
9.144 E-01 



MASS and MOMENT OF INERTIA 

kilogram-force second squared 
per meter (kgf ■ s 2 /m) 

pound foot squared (lb ■ ft ) 

pound inch squared (lb ■ in. 2 ) 

ton, metric (t) 

ton, short (2,000 lb) 



kilogram (kg) 

kilogram meter squared (kg ■ m ) 

kilogram meter squared (kg ■ m 2 ) 

kilogram (kg) 

kilogram (kg) 



9.806 65 E+00 
4.214 011 E-02 
2.926 397 E-04 
1.0E+03 
9.071 847 E+02 



MASS DIVIDED BY AREA 

pound per square foot (lb/ft ) 

pound per square inch 
(not pound force) (lb/in. ) 



kilogram per square meter (kg/m ) 4.882 428 E+00 



kilogram per square meter (kg/m ) 7.030 696 E+02 



MASS DIVIDED BY LENGTH 

pound per foot (lb/ft) 
pound per inch (lb/in.) 
pound per yard (lb/yd) 



kilogram per meter (kg/m) 
kilogram per meter (kg/m) 
kilogram per meter (kg/m) 



1.488 164 E+00 
1.785 797 E+01 
4.960 546 E-01 



PRESSURE or STRESS (FORCE DIVIDED BY AREA) 

kilogram-force per square centimeter (kgf/cm 2 ) pascal (Pa) 



kilogram-force per square meter (kgf/m 2 ) 



pascal (Pa) 



kilogram-force per square millimeter (kgf/mm ) pascal (Pa) 

kip per square inch (ksi) (kip/in. ) pascal (Pa) 

kip per square inch (ksi) (kip/in. ) kilopascal (kPa) 

pound-force per square foot (lbf/ft ) pascal (Pa) 

pound-force per square inch (psi) (lbf/in. ) pascal (Pa) 

pound-force per square inch (psi) (lbf/in. 2 ) kilopascal (kPa) 

psi (pound-force per square inch) (lbf/in. ) pascal (Pa) 

psi (pound-force per square inch) (lbf/in. 2 ) kilopascal (kPa) 



9.806 65 E+04 
9.806 65 E+00 
9.806 65 E+06 
6.894 757 E+06 
6.894 757 E+03 
4.788 026 E+01 
6.894 757 E+03 
6.894 757 E+00 
6.894 757 E+03 
6.894 757 E+00 



xvi 



NIST NCSTAR 1-7, WTC Investigation 



Metric Conversion Table 



To convert from 



to 



Multiply by 



TEMPERATURE 

degree Celsius (°C) 
degree centigrade 
degree Fahrenheit (°F) 
degree Fahrenheit (°F) 
kelvin (K) 



kelvin (K) 

degree Celsius (°C) 

degree Celsius (°C) 

kelvin (K) 

degree Celsius (°C) 



T/K = t/°C + 273.15 
t/°C » t /deg. cent. 
t/°C = (t/°F-32)/1.8 
T/K=(t/°F+459.67)/l. 
t/°C = T/K 2 273.15 



TEMPERATURE INTERVAL 

degree Celsius (°C) 
degree centigrade 
degree Fahrenheit (°F) 
degree Fahrenheit (°F) 
degree Rankine (°R) 



kelvin (K) 
degree Celsius (°C) 
degree Celsius (°C) 
kelvin (K) 
kelvin (K) 



1.0E+00 
1.0E+00 
5.555 556 E-01 
5.555 556 E-01 
5.555 556 E-01 



VELOCITY (includes SPEED) 

foot per second (ft/s) 
inch per second (in./s) 
kilometer per hour (km/h) 
mile per hour (mi/h) 
mile per minute (mi/min) 



meter per second (m/s) 
meter per second (m/s) 
meter per second (m/s) 
kilometer per hour (km/h) 
meter per second (m/s) 



3.048 E-01 
2.54 E-02 
2.777 778 E-01 
1.609 344 E+00 
2.682 24 E+01 



VOLUME (includes CAPACITY) 

cubic foot (ft 3 ) 
cubic inch (in. ) 
cubic yard (yd 3 ) 
gallon (U.S.) (gal) 
gallon (U.S.) (gal) 
liter (L) 

ounce (U.S. fluid) (fl oz) 
ounce (U.S. fluid) (fl oz) 



cubic meter (m ) 
cubic meter (m ) 
cubic meter (m 3 ) 
cubic meter (m 3 ) 
liter (L) 

cubic meter (m 3 ) 
cubic meter (m ) 
milliliter (mL) 



2.831 685 E-02 
1.638 706 E-05 
7.645 549 E-01 
3.785 412 E-03 
3.785 412 E+00 
1.0 E-03 
2.957 353 E-05 
2.957 353 E+01 



NISTNCSTAR 1-7, WTC Investigation 



xvn 



Metric Conversion Table 



This page intentionally left blank. 



xviii NIST NCSTAR 1-7, WTC Investigation 



Preface 



Genesis of This Investigation 

Immediately following the terrorist attack on the World Trade Center (WTC) on September 11, 2001, the 
Federal Emergency Management Agency (FEMA) and the American Society of Civil Engineers began 
planning a building performance study of the disaster. The week of October 7, as soon as the rescue and 
search efforts ceased, the Building Performance Study Team went to the site and began its assessment. 
This was to be a brief effort, as the study team consisted of experts who largely volunteered their time 
away from their other professional commitments. The Building Performance Study Team issued its 
report in May 2002, fulfilling its goal "to determine probable failure mechanisms and to identify areas of 
future investigation that could lead to practical measures for improving the damage resistance of buildings 
against such unforeseen events." 

On August 21, 2002, with funding from the U.S. Congress through FEMA, the National Institute of 
Standards and Technology (NIST) announced its building and fire safety investigation of the WTC 
disaster. On October 1, 2002, the National Construction Safety Team Act (Public Law 107-231), was 
signed into law. The NIST WTC Investigation was conducted under the authority of the National 
Construction Safety Team Act. 

The goals of the investigation of the WTC disaster were: 

• To investigate the building construction, the materials used, and the technical conditions that 
contributed to the outcome of the WTC disaster. 

• To serve as the basis for: 

- Improvements in the way buildings are designed, constructed, maintained, and used; 

- Improved tools and guidance for industry and safety officials; 

- Recommended revisions to current codes, standards, and practices; and 

- Improved public safety. 

The specific objectives were: 

1. Determine why and how WTC 1 and WTC 2 collapsed following the initial impacts of the 
aircraft and why and how WTC 7 collapsed; 

2. Determine why the injuries and fatalities were so high or low depending on location, 
including all technical aspects of fire protection, occupant behavior, evacuation, and 
emergency response; 

3. Determine what procedures and practices were used in the design, construction, operation, 
and maintenance of WTC 1, 2, and 7; and 

4. Identify, as specifically as possible, areas in current building and fire codes, standards, and 
practices that warrant revision. 



NISTNCSTAR 1-7, WTC Investigation 



Preface 



NIST is a nonregulatory agency of the U.S. Department of Commerce's Technology Administration. The 
purpose of NIST investigations is to improve the safety and structural integrity of buildings in the United 
States, and the focus is on fact finding. NIST investigative teams are authorized to assess building 
performance and emergency response and evacuation procedures in the wake of any building failure that 
has resulted in substantial loss of life or that posed significant potential of substantial loss of life. NIST 
does not have the statutory authority to make findings of fault nor negligence by individuals or 
organizations. Further, no part of any report resulting from a NIST investigation into a building failure or 
from an investigation under the National Construction Safety Team Act may be used in any suit or action 
for damages arising out of any matter mentioned in such report (15 USC 281a, as amended by Public 
Law 107-231). 

Organization of the Investigation 

The National Construction Safety Team for this Investigation, appointed by the then NIST Director, 
Dr. Arden L. Bement, Jr., was led by Dr. S. Shyam Sunder. Dr. William L. Grosshandler served as 
Associate Lead Investigator, Mr. Stephen A. Cauffman served as Program Manager for Administration, 
and Mr. Harold E. Nelson served on the team as a private sector expert. The Investigation included eight 
interdependent projects whose leaders comprised the remainder of the team. A detailed description of 
each of these eight projects is available at http://wtc.nist.gov. The purpose of each project is summarized 
in Table P-l, and the key interdependencies among the projects are illustrated in Fig. P-l. 



Table P-l. Federal building and fire safety investigation of the WTC disaster. 



Technical Area and Project Leader 


Project Purpose 


Analysis of Building and Fire Codes and 
Practices; Project Leaders: Dr. H. S. Lew 
and Mr. Richard W. Bukowski 


Document and analyze the code provisions, procedures, and 
practices used in the design, construction, operation, and 
maintenance of the structural, passive fire protection, and 
emergency access and evacuation systems of WTC 1, 2, and 7. 


Baseline Structural Performance and 
Aircraft Impact Damage Analysis; Project 
Leader: Dr. Fahim H. Sadek 


Analyze the baseline performance of WTC 1 and WTC 2 under 
design, service, and abnormal loads, and aircraft impact damage on 
the structural, fire protection, and egress systems. 


Mechanical and Metallurgical Analysis of 
Structural Steel; Project Leader: Dr. Frank 
W. Gayle 


Determine and analyze the mechanical and metallurgical properties 
and quality of steel, weldments, and connections from steel 
recovered from WTC 1, 2, and 7. 


Investigation of Active Fire Protection 
Systems; Project Leader: Dr. David 
D. Evans; Dr. William Grosshandler 


Investigate the performance of the active fire protection systems in 
WTC 1 , 2, and 7 and their role in fire control, emergency response, 
and fate of occupants and responders. 


Reconstruction of Thermal and Tenability 
Environment; Project Leader: Dr. Richard 
G. Gann 


Reconstruct the time-evolving temperature, thermal environment, 
and smoke movement in WTC 1 , 2, and 7 for use in evaluating the 
structural performance of the buildings and behavior and fate of 
occupants and responders. 


Structural Fire Response and Collapse 
Analysis; Project Leaders: Dr. John 
L. Gross and Dr. Therese P. McAllister 


Analyze the response of the WTC towers to fires with and without 
aircraft damage, the response of WTC 7 in fires, the performance 
of composite steel-trussed floor systems, and determine the most 
probable structural collapse sequence for WTC 1, 2, and 7. 


Occupant Behavior, Egress, and Emergency 
Communications; Project Leader: Mr. Jason 
D. Averill 


Analyze the behavior and fate of occupants and responders, both 
those who survived and those who did not, and the performance of 
the evacuation system. 


Emergency Response Technologies and 
Guidelines; Project Leader: Mr. J. Randall 
Law son 


Document the activities of the emergency responders from the time 
of the terrorist attacks on WTC 1 and WTC 2 until the collapse of 
WTC 7, including practices followed and technologies used. 



NIST NCSTAR 1-7, WTC Investigation 



Preface 



WTC Building 
Performance Study 
Recommendations 

Government, 
Industry, 
Professional, 
Academic Inp 

Public Inputs 



Video/ 

Photographic 

Records 

Oral History Data 

Emergency 

Response 

Records 

Recovered 
Structural Steel/ 



fMisr 



NIST WTC Investigation Projects 



, / Analysis of V^^ 
\ I Steel 1 


■■n^^^/ Structural \ 
V Collapse J 






/ Baseline 








/ Performance 








V & Impact 
\ Damage . 


/ [ Thermal and \ 
I Tenability J 
\ Environment J 






/ Analysis of \ . 
I Codes and ) / 
V Practices J 1 


Active Fire \ V 
Protection / 3- -\ 

^ ^ ( Emergency \ 

I Response / 


1 Evacuation 1 













Figure P-l. The eight projects in the federal building and fire safety 
investigation of the WTC disaster. 



National Construction Safety Team Advisory Committee 

The NIST Director also established an advisory committee as mandated under the National Construction 
Safety Team Act. The initial members of the committee were appointed following a public solicitation. 
These were: 



Paul Fitzgerald, Executive Vice President (retired) FM Global, National Construction Safety 
Team Advisory Committee Chair 

John Barsom, President, Barsom Consulting, Ltd. 

John Bryan, Professor Emeritus, University of Maryland 

David Collins, President, The Preview Group, Inc. 

Glenn Corbett, Professor, John Jay College of Criminal Justice 

Philip DiNenno, President, Hughes Associates, Inc. 



NISTNCSTAR 1-7, WTC Investigation 



xxi 



• 



• 



• 



Preface 



Robert Hanson, Professor Emeritus, University of Michigan 

Charles Thornton, Co-Chairman and Managing Principal, The Thornton-Tomasetti Group, 
Inc. 

Kathleen Tierney, Director, Natural Hazards Research and Applications Information Center, 
University of Colorado at Boulder 

Forman Williams, Director, Center for Energy Research, University of California at San 
Diego 

This National Construction Safety Team Advisory Committee provided technical advice during the 
Investigation and commentary on drafts of the Investigation reports prior to their public release. NIST 
has benefited from the work of many people in the preparation of these reports, including the National 
Construction Safety Team Advisory Committee. The content of the reports and recommendations, 
however, are solely the responsibility of NIST. 

Public Outreach 

During the course of this Investigation, NIST held public briefings and meetings (listed in Table P-2) to 
solicit input from the public, present preliminary findings, and obtain comments on the direction and 
progress of the Investigation from the public and the Advisory Committee. 

NIST maintained a publicly accessible Web site during this Investigation at http://wtc.nist.gov. The site 
contained extensive information on the background and progress of the Investigation. 

NIST's WTC Public-Private Response Plan 

The collapse of the WTC buildings has led to broad reexamination of how tall buildings are designed, 
constructed, maintained, and used, especially with regard to major events such as fires, natural disasters, 
and terrorist attacks. Reflecting the enhanced interest in effecting necessary change, NIST, with support 
from Congress and the Administration, has put in place a program, the goal of which is to develop and 
implement the standards, technology, and practices needed for cost-effective improvements to the safety 
and security of buildings and building occupants, including evacuation, emergency response procedures, 
and threat mitigation. 

The strategy to meet this goal is a three-part NIST-led public-private response program that includes: 

• A federal building and fire safety investigation to study the most probable factors that 
contributed to post-aircraft impact collapse of the WTC towers and the 47-story WTC 7 
building, and the associated evacuation and emergency response experience. 

• A research and development (R&D) program to (a) facilitate the implementation of 
recommendations resulting from the WTC Investigation, and (b) provide the technical basis 
for cost-effective improvements to national building and fire codes, standards, and practices 
that enhance the safety of buildings, their occupants, and emergency responders. 



xxii NIST NCSTAR 1-7, WTC Investigation 



Preface 



Table P-2. Public meetings and briefings of the WTC Investigation. 


Date 


Location 


Principal Agenda 


June 24, 2002 


New York City, NY 


Public meeting: Public comments on the Draft Plan for the 
pending WTC Investigation. 


August 21, 2002 


Gaithersburg, MD 


Media briefing announcing the formal start of the Investigation. 


December 9, 2002 


Washington, DC 


Media briefing on release of the Public Update and NIST request 
for photographs and videos. 


April 8, 2003 


New York City, NY 


Joint public forum with Columbia University on first-person 
interviews. 


April 29-30, 2003 


Gaithersburg, MD 


NCST Advisory Committee meeting on plan for and progress on 
WTC Investigation with a public comment session. 


May 7, 2003 


New York City, NY 


Media briefing on release of May 2003 Progress Report. 


August 26-27, 2003 


Gaithersburg, MD 


NCST Advisory Committee meeting on status of the WTC 
investigation with a public comment session. 


September 17,2003 


New York City, NY 


Media and public briefing on initiation of first-person data 
collection projects. 


December 2-3, 2003 


Gaithersburg, MD 


NCST Advisory Committee meeting on status and initial results 
and release of the Public Update with a public comment session. 


February 12,2004 


New York City, NY 


Public meeting on progress and preliminary findings with public 
comments on issues to be considered in formulating final 
recommendations. 


June 18,2004 


New York City, NY 


Media/public briefing on release of June 2004 Progress Report. 


June 22-23, 2004 


Gaithersburg, MD 


NCST Advisory Committee meeting on the status of and 
preliminary findings from the WTC Investigation with a public 
comment session. 


August 24, 2004 


Northbrook, IL 


Public viewing of standard fire resistance test of WTC floor 
system at Underwriters Laboratories, Inc. 


October 19-20, 2004 


Gaithersburg, MD 


NCST Advisory Committee meeting on status and near complete 
set of preliminary findings with a public comment session. 


November 22, 2004 


Gaithersburg, MD 


NCST Advisory Committee discussion on draft annual report to 
Congress, a public comment session, and a closed session to 
discuss pre-draft recommendations for WTC Investigation. 


April 5, 2005 


New York City, NY 


Media and public briefing on release of the probable collapse 
sequence for the WTC towers and draft reports for the projects on 
codes and practices, evacuation, and emergency response. 


June 23, 2005 


New York City, NY 


Media and public briefing on release of all draft reports for the 
WTC towers and draft recommendations for public comment. 


September 12-13, 
2005 


Gaithersburg, MD 


NCST Advisory Committee meeting on disposition of public 
comments and update to draft reports for the WTC towers. 


September 13-15, 
2005 


Gaithersburg, MD 


WTC Technical Conference for stakeholders and technical 
community for dissemination of findings and recommendations 
and opportunity for public to make technical comments. 



• A dissemination and technical assistance program (DTAP) to (a) engage leaders of the 
construction and building community in ensuring timely adoption and widespread use of 
proposed changes to practices, standards, and codes resulting from the WTC Investigation 
and the R&D program, and (b) provide practical guidance and tools to better prepare facility 
owners, contractors, architects, engineers, emergency responders, and regulatory authorities 
to respond to future disasters. 

The desired outcomes are to make buildings, occupants, and first responders safer in future disaster 
events. 



NISTNCSTAR 1-7, WTC Investigation 



xxm 



Preface 



National Construction Safety Team Reports on the WTC Investigation 

A final report on the collapse of the WTC towers is being issued as NIST NCSTAR 1. A companion 
report on the collapse of WTC 7 is being issued as NIST NCSTAR 1A. The present report is one of a set 
that provides more detailed documentation of the Investigation findings and the means by which these 
technical results were achieved. As such, it is part of the archival record of this Investigation. The titles 
of the full set of Investigation publications are: 

NIST (National Institute of Standards and Technology). 2005. Federal Building and Fire Safety 
Investigation of the World Trade Center Disaster: Final Report on the Collapse of the World Trade 
Center Towers. NIST NCSTAR 1. Gaithersburg, MD, September. 

NIST (National Institute of Standards and Technology). 2006. Federal Building and Fire Safety 
Investigation of the World Trade Center Disaster: Final Report on the Collapse of World Trade Center 7. 
NIST NCSTAR 1A. Gaithersburg, MD. 

Lew, H. S., R. W. Bukowski, and N. J. Carino. 2005. Federal Building and Fire Safety Investigation of 
the World Trade Center Disaster: Design, Construction, and Maintenance of Structural and Life Safety 
Systems. NIST NCSTAR 1-1. National Institute of Standards and Technology. Gaithersburg, MD, 
September. 

Fanella, D. A., A. T. Derecho, and S. K. Ghosh. 2005. Federal Building and Fire Safety 
Investigation of the World Trade Center Disaster: Design and Construction of Structural Systems. 
NIST NCSTAR 1-1 A. National Institute of Standards and Technology. Gaithersburg, MD, 
September. 

Ghosh, S. K., and X. Liang. 2005. Federal Building and Fire Safety Investigation of the World 
Trade Center Disaster: Comparison of Building Code Structural Requirements. NIST 
NCSTAR 1-1B. National Institute of Standards and Technology. Gaithersburg, MD, September. 

Fanella, D. A., A. T. Derecho, and S. K. Ghosh. 2005. Federal Building and Fire Safety 
Investigation of the World Trade Center Disaster: Maintenance and Modifications to Structural 
Systems. NIST NCSTAR 1-lC. National Institute of Standards and Technology. Gaithersburg, 
MD, September. 

Grill, R. A., and D. A. Johnson. 2005. Federal Building and Fire Safety Investigation of the World 
Trade Center Disaster: Fire Protection and Life Safety Provisions Applied to the Design and 
Construction of World Trade Center 1, 2, and 7 and Post-Construction Provisions Applied after 
Occupancy. NIST NCSTAR 1-1D. National Institute of Standards and Technology. Gaithersburg, 
MD, September. 

Razza, J. C, and R. A. Grill. 2005. Federal Building and Fire Safety Investigation of the World 
Trade Center Disaster: Comparison of Codes, Standards, and Practices in Use at the Time of the 
Design and Construction of World Trade Center 1, 2, and 7. NIST NCSTAR 1-1E. National 
Institute of Standards and Technology. Gaithersburg, MD, September. 

Grill, R. A., D. A. Johnson, and D. A. Fanella. 2005. Federal Building and Fire Safety 
Investigation of the World Trade Center Disaster: Comparison of the 1968 and Current (2003) New 



xxiv NIST NCSTAR 1-7, WTC Investigation 



Preface 



York City Building Code Provisions. NIST NCSTAR 1-1F. National Institute of Standards and 
Technology. Gaithersburg, MD, September. 

Grill, R. A., and D. A. Johnson. 2005. Federal Building and Fire Safety Investigation of the World 
Trade Center Disaster: Amendments to the Fire Protection and Life Safety Provisions of the New 
York City Building Code by Local Laws Adopted While World Trade Center 1, 2, and 7 Were in 
Use. NIST NCSTAR 1-1G. National Institute of Standards and Technology. Gaithersburg, MD, 
September. 

Grill, R. A., and D. A. Johnson. 2005. Federal Building and Fire Safety Investigation of the World 
Trade Center Disaster: Post-Construction Modifications to Fire Protection and Life Safety Systems 
of World Trade Center 1 and 2. NIST NCSTAR 1-1H. National Institute of Standards and 
Technology. Gaithersburg, MD, September. 

Grill, R. A., D. A. Johnson, and D. A. Fanella. 2005. Federal Building and Fire Safety Investigation 
of the World Trade Center Disaster: Post-Construction Modifications to Fire Protection, Life 
Safety, and Structural Systems of World Trade Center 7. NIST NCSTAR 1-1 1. National Institute of 
Standards and Technology. Gaithersburg, MD, September. 

Grill, R. A., and D. A. Johnson. 2005. Federal Building and Fire Safety Investigation of the World 
Trade Center Disaster: Design, Installation, and Operation of Fuel System for Emergency Power in 
World Trade Center 7. NIST NCSTAR 1-1 J. National Institute of Standards and Technology. 
Gaithersburg, MD, September. 

Sadek, F. 2005. Federal Building and Fire Safety Investigation of the World Trade Center Disaster: 
Baseline Structural Performance and Aircraft Impact Damage Analysis of the World Trade Center 
Towers. NIST NCSTAR 1-2. National Institute of Standards and Technology. Gaithersburg, MD, 
September. 

Faschan, W. J., and R. B. Garlock. 2005. Federal Building and Fire Safety Investigation of the 
World Trade Center Disaster: Reference Structural Models and Baseline Performance Analysis of 
the World Trade Center Towers. NIST NCSTAR 1-2A. National Institute of Standards and 
Technology. Gaithersburg, MD, September. 

Kirkpatrick, S. W., R. T. Bocchieri, F. Sadek, R. A. MacNeill, S. Holmes, B. D. Peterson, 
R. W. Cilke, C. Navarro. 2005. Federal Building and Fire Safety Investigation of the World Trade 
Center Disaster: Analysis of Aircraft Impacts into the World Trade Center Towers, NIST 
NCSTAR 1-2B. National Institute of Standards and Technology. Gaithersburg, MD, September. 

Gayle, F. W., R. J. Fields, W. E. Luecke, S. W. Banovic, T. Foecke, C. N. McCowan, T. A. Siewert, and 
J. D. McColskey. 2005. Federal Building and Fire Safety Investigation of the World Trade Center 
Disaster: Mechanical and Metallurgical Analysis of Structural Steel. NIST NCSTAR 1-3. National 
Institute of Standards and Technology. Gaithersburg, MD, September. 

Luecke, W. E., T. A. Siewert, and F. W. Gayle. 2005. Federal Building and Fire Safety 
Investigation of the World Trade Center Disaster: Contemporaneous Structural Steel 
Specifications. NIST Special Publication 1-3 A. National Institute of Standards and Technology. 
Gaithersburg, MD, September. 



NIST NCSTAR 1-7, WTC Investigation 



Preface 



Banovic, S. W. 2005. Federal Building and Fire Safety Investigation of the World Trade Center 
Disaster: Steel Inventory and Identification. NIST NCSTAR 1-3B. National Institute of Standards 
and Technology. Gaithersburg, MD, September. 

Banovic, S. W., and T. Foecke. 2005. Federal Building and Fire Safety Investigation of the World 
Trade Center Disaster: Damage and Failure Modes of Structural Steel Components. NIST 
NCSTAR 1-3C. National Institute of Standards and Technology. Gaithersburg, MD, September. 

Luecke, W. E., J. D. McColskey, C. N. McCowan, S. W. Banovic, R. J. Fields, T. Foecke, 
T. A. Siewert, and F. W. Gayle. 2005. Federal Building and Fire Safety Investigation of the World 
Trade Center Disaster: Mechanical Properties of Structural Steels. NIST NCSTAR 1-3D. 
National Institute of Standards and Technology. Gaithersburg, MD, September. 

Banovic, S. W., C. N. McCowan, and W. E. Luecke. 2005. Federal Building and Fire Safety 
Investigation of the World Trade Center Disaster: Physical Properties of Structural Steels. NIST 
NCSTAR 1-3E. National Institute of Standards and Technology. Gaithersburg, MD, September. 

Evans, D. D., R. D. Peacock, E. D. Kuligowski, W. S. Dols, and W. L. Grosshandler. 2005. Federal 
Building and Fire Safety Investigation of the World Trade Center Disaster: Active Fire Protection 
Systems. NIST NCSTAR 1-4. National Institute of Standards and Technology. Gaithersburg, MD, 
September. 

Kuligowski, E. D., D. D. Evans, and R. D. Peacock. 2005. Federal Building and Fire Safety 
Investigation of the World Trade Center Disaster: Post-Construction Fires Prior to September 11, 
2001. NIST NCSTAR 1-4A. National Institute of Standards and Technology. Gaithersburg, MD, 
September. 

Hopkins, M., J. Schoenrock, and E. Budnick. 2005. Federal Building and Fire Safety Investigation 
of the World Trade Center Disaster: Fire Suppression Systems. NIST NCSTAR 1-4B. National 
Institute of Standards and Technology. Gaithersburg, MD, September. 

Keough, R. J., and R. A. Grill. 2005. Federal Building and Fire Safety Investigation of the World 
Trade Center Disaster: Fire Alarm Systems. NIST NCSTAR 1-4C. National Institute of Standards 
and Technology. Gaithersburg, MD, September. 

Ferreira, M. J., and S. M. Strege. 2005. Federal Building and Fire Safety Investigation of the 
World Trade Center Disaster: Smoke Management Systems. NIST NCSTAR 1-4D. National 
Institute of Standards and Technology. Gaithersburg, MD, September. 

Gann, R. G., A. Hamins, K. B. McGrattan, G. W. Mulholland, H. E. Nelson, T. J. Ohlemiller, 
W. M. Pitts, and K. R. Prasad. 2005. Federal Building and Fire Safety Investigation of the World Trade 
Center Disaster: Reconstruction of the Fires in the World Trade Center Towers. NIST NCSTAR 1-5. 
National Institute of Standards and Technology. Gaithersburg, MD, September. 

Pitts, W. M., K. M. Butler, and V. Junker. 2005. Federal Building and Fire Safety Investigation of 
the World Trade Center Disaster: Visual Evidence, Damage Estimates, and Timeline Analysis. 
NIST NCSTAR 1-5 A. National Institute of Standards and Technology. Gaithersburg, MD, 
September. 



xxvi NIST NCSTAR 1-7, WTC Investigation 



Preface 



Hamins, A., A. Maranghides, K. B. McGrattan, E. Johnsson, T. J. Ohlemiller, M. Donnelly, 
J. Yang, G. Mulholland, K. R. Prasad, S. Kukuck, R. Anleitner and T. McAllister. 2005. Federal 
Building and Fire Safety Investigation of the World Trade Center Disaster: Experiments and 
Modeling of Structural Steel Elements Exposed to Fire. NIST NCSTAR 1-5B. National Institute of 
Standards and Technology. Gaithersburg, MD, September. 

Ohlemiller, T. J., G. W. Mulholland, A. Maranghides, J. J. Filliben, and R. G. Gann. 2005. Federal 
Building and Fire Safety Investigation of the World Trade Center Disaster: Fire Tests of Single 
Office Workstations. NIST NCSTAR 1-5C. National Institute of Standards and Technology. 
Gaithersburg, MD, September. 

Gann, R. G., M. A. Riley, J. M. Repp, A. S. Whittaker, A. M. Reinhorn, and P. A. Hough. 2005. 
Federal Building and Fire Safety Investigation of the World Trade Center Disaster: Reaction of 
Ceiling Tile Systems to Shocks. NIST NCSTAR 1-5D. National Institute of Standards and 
Technology. Gaithersburg, MD, September. 

Hamins, A., A. Maranghides, K. B. McGrattan, T. J. Ohlemiller, and R. Anleitner. 2005. Federal 
Building and Fire Safety Investigation of the World Trade Center Disaster: Experiments and 
Modeling of Multiple Workstations Burning in a Compartment. NIST NCSTAR 1-5E. National 
Institute of Standards and Technology. Gaithersburg, MD, September. 

McGrattan, K. B., C. Bouldin, and G. Forney. 2005. Federal Building and Fire Safety 
Investigation of the World Trade Center Disaster: Computer Simulation of the Fires in the World 
Trade Center Towers. NIST NCSTAR 1-5F. National Institute of Standards and Technology. 
Gaithersburg, MD, September. 

Prasad, K. R., and H. R. Baum. 2005. Federal Building and Fire Safety Investigation of the World 
Trade Center Disaster: Fire Structure Interface and Thermal Response of the World Trade Center 
Towers. NIST NCSTAR 1-5G. National Institute of Standards and Technology. Gaithersburg, 
MD, September. 

Gross, J. L., and T. McAllister. 2005. Federal Building and Fire Safety Investigation of the World Trade 
Center Disaster: Structural Fire Response and Probable Collapse Sequence of the World Trade Center 
Towers. NIST NCSTAR 1-6. National Institute of Standards and Technology. Gaithersburg, MD, 
September. 

Carino, N. J., M. A. Starnes, J. L. Gross, J. C. Yang, S. Kukuck, K. R. Prasad, and R. W. Bukowski. 
2005. Federal Building and Fire Safety Investigation of the World Trade Center Disaster: Passive 
Fire Protection. NIST NCSTAR 1-6A. National Institute of Standards and Technology. 
Gaithersburg, MD, September. 

Gross, J., F. Hervey, M. Izydorek, J. Mammoser, and J. Treadway. 2005. Federal Building and 
Fire Safety Investigation of the World Trade Center Disaster: Fire Resistance Tests of Floor Truss 
Systems. NIST NCSTAR 1-6B. National Institute of Standards and Technology. Gaithersburg, 
MD, September. 

Zarghamee, M. S., S. Bolourchi, D. W. Eggers, O. O. Erbay, F. W. Kan, Y. Kitane, A. A. Liepins, 
M. Mudlock, W. I. Naguib, R. P. Ojdrovic, A. T. Sarawit, P. R Barrett, J. L. Gross, and 



NIST NCSTAR 1-7, WTC Investigation xxvii 



Preface 



T. P. McAllister. 2005. Federal Building and Fire Safety Investigation of the World Trade Center 
Disaster: Component, Connection, and Subsystem Structural Analysis. NIST NCSTAR 1-6C. 
National Institute of Standards and Technology. Gaithersburg, MD, September. 

Zarghamee, M. S., Y. Kitane, O. O. Erbay, T. P. McAllister, and J. L. Gross. 2005. Federal 
Building and Fire Safety Investigation of the World Trade Center Disaster: Global Structural 
Analysis of the Response of the World Trade Center Towers to Impact Damage and Fire. NIST 
NCSTAR 1-6D. National Institute of Standards and Technology. Gaithersburg, MD, September. 

McAllister, T., R. W. Bukowski, R. G. Gann, J. L. Gross, K. B. McGrattan, H. E. Nelson, L. Phan, 
W. M. Pitts, K. R. Prasad, F. Sadek. 2006. Federal Building and Fire Safety Investigation of the World 
Trade Center Disaster: Structural Fire Response and Probable Collapse Sequence of World Trade 
Center 7. (Provisional). NIST NCSTAR 1-6E. National Institute of Standards and Technology. 
Gaithersburg, MD. 

Gilsanz, R., V. Arbitrio, C. Anders, D. Chlebus, K. Ezzeldin, W. Guo, P. Moloney, A. Montalva, 
J. Oh, K. Rubenacker. 2006. Federal Building and Fire Safety Investigation of the World Trade 
Center Disaster: Structural Analysis of the Response of World Trade Center 7 to Debris Damage 
and Fire. (Provisional). NIST NCSTAR 1-6F. National Institute of Standards and Technology. 
Gaithersburg, MD. 

Kim, W. 2006. Federal Building and Fire Safety Investigation of the World Trade Center 
Disaster: Analysis of September 11, 2001, Seismogram Data. (Provisional). NIST NCSTAR 1-6G. 
National Institute of Standards and Technology. Gaithersburg, MD. 

Nelson, K. 2006. Federal Building and Fire Safety Investigation of the World Trade Center 
Disaster: The Con Ed Substation in World Trade Center 7. (Provisional). NIST NCSTAR 1-6H. 
National Institute of Standards and Technology. Gaithersburg, MD. 

Averill, J. D., D. S. Mileti, R. D. Peacock, E. D. Kuligowski, N. Groner, G. Proulx, P. A. Reneke, and 
H. E. Nelson. 2005. Federal Building and Fire Safety Investigation of the World Trade Center Disaster: 
Occupant Behavior, Egress, and Emergency Communication. NIST NCSTAR 1-7. National Institute of 
Standards and Technology. Gaithersburg, MD, September. 

Fahy, R., and G. Proulx. 2005. Federal Building and Fire Safety Investigation of the World Trade 
Center Disaster: Analysis of Published Accounts of the World Trade Center Evacuation. NIST 
NCSTAR 1-7A. National Institute of Standards and Technology. Gaithersburg, MD, September. 

Zmud, J. 2005. Federal Building and Fire Safety Investigation of the World Trade Center 
Disaster: Technical Documentation for Survey Administration. NIST NCSTAR 1-7B. National 
Institute of Standards and Technology. Gaithersburg, MD, September. 

Lawson, J. R., and R. L. Vettori. 2005. Federal Building and Fire Safety Investigation of the World 
Trade Center Disaster: The Emergency Response Operations. NIST NCSTAR 1-8. National Institute of 
Standards and Technology. Gaithersburg, MD, September. 



xxviii NIST NCSTAR 1-7, WTC Investigation 



Acknowledgments 



The authors wish to acknowledge the contributions of a number of people, without whom, this project 
would not have been successful. 

Robyn Gershon, Erin Hogan, and Stephen Morse, Columbia University; Nora Marshall and others from 
the National Transportation Safety Board; Henry Quarantelli, and Benigno Aguirre from the Disaster 
Research Center, University of Delaware; Linda Bourque - University of California - Los Angeles; 
Richard Mendelson, Occupational Safety and Health Administration; Jake Pauls; and Edwina Juillet all 
provided professional insight into issues, procedures, and pitfalls of conducting research involving human 
behavior in disasters. 

Rita Fahy, National Fire Protection Association analyzed media accounts of the September 11, 2001, 
attacks and provided the EXIT89 model and support in its use. 

Jay Cohen and George Duke, employees of the Corporation Counsel of the City of New York, provided 
access to emergency communications and 9-1-1 records that provided details of the experiences and 
conditions above the floors of impact in the WTC towers. 

Frank Lombardi, Saroj Bhol, Alan Reiss, and Nancy Seliga, PANYNJ provided access to numerous Port 
Authority records on the World Trade Center egress system and building procedures and facilitated access 
to Port Authority personnel with extensive knowledge of the building operation and the events of 
September 11,2001. 

Patti Adler, Julie Gailus, Janet Jacobs, and Lori Peek - University of Colorado - Boulder; Martha van 
Haitsma and Virginia Bartot - University of Chicago provided guidance on the analysis of qualitative 
interview data. 

Peter Thompson, Integrated Environmental Solutions, Inc.; Ed Galea and Steven Gwynne, University of 
Greenwich - UK; and Jeremy Fraser-Mitchell, BRE - FRS, UK provided support and suggestions on 
appropriate techniques for modeling evacuation in high-rise structures. 

Finally, National Institute of Standards and Technology would like to gratefully acknowledge the more 
than 1,000 survivors and families of victims of the September 11, 2001, attacks that generously agreed to 
be interviewed about their difficult and often tragic experiences on September 11, 2001. Their detailed 
accounts provide a continuing memory of the attacks and the victims of September 11, 2001. 



NIST NCSTAR 1-7, WTC Investigation xxix 



Acknowledgments 



This page intentionally left blank. 



NIST NCSTAR 1-7, WTC Investigation 



Executive Summary 



e.i OVERVIEW 

While most attention has properly focused on the nearly three thousand people who lost their lives at the 
World Trade Center (WTC) site on September 11, 2001, five times that many people successfully 
evacuated from the WTC towers due to heroic efforts of occupants, as well as emergency responders. 
Understanding why many, yet not all, survived the WTC attacks was one of the four objectives of the 
Federal building and fire safety investigation of the WTC disaster led by the National Institute of 
Standards and Technology (NIST). 

Success in evacuating a building in an emergency can be characterized by two quantities: the time people 
needed to evacuate and the time available for them to do so. To the extent the first time exceeded the 
second, it follows that there will be casualties. When the second time exceeds the first, perhaps by some 
suitable margin, nearly all should be able to evacuate the building. 

For the WTC towers, the times available for escape were cataclysmically established by the collapses of 
the buildings. Those times were not known in advance by the building occupants or the responders. The 
times were also considerably shorter, by a factor of three or four, than the time needed to clear the tenant 
spaces of WTC 1 following the 1993 bombing and an additional factor of two shorter than the time 
needed to clear the last person from the elevators in the building. Further, some occupants would have 
been unable to evacuate the buildings given any amount of time due to injuries, entrapment, and/or toxic 
exposure. 

NIST examined the design of the building, the behavior of the people, and the evacuation process in detail 
to ascertain the factors that factored prominently in the time needed for evacuation. 

In order to accomplish this objective, numerous sources of data were collected and analyzed, including: 
over 1,000 new interviews with survivors; a collection of over 700 published interviews with 
WTC survivors; 9-1-1 emergency calls; transcripts of emergency communication among building 
personnel and emergency responders; historical building design drawings, memoranda, and calculations; 
building modifications and upgrades; formal complaints filed with Occupational Safety and Health 
Administration; and other relevant material. 

There were three forms of interviews with survivors: 803 telephone interviews, over 225 face-to-face 
interviews, and 6 focus groups. The telephone interviewees were randomly selected using independent 
proportionate stratification from a list of occupants who had badges to enter WTC 1 or WTC 2 on 
September 1 1 , 200 1 . In other words, each occupant of a particular tower had an equal probability of 
being selected. Roughly 400 occupants in each tower were interviewed in order to achieve a high level of 
statistical precision within each tower. Reported percentages from tower-specific survey data (n=400) 
exhibited sampling errors no greater than 2.5 percentage points, and 95 percent confidence intervals of 
percentages are no greater than ± 5 percentage points. This level of precision was more than adequate for 
examining characteristics of occupants and egress attributes. With telephone interview results, primary 
statistical analyses were in the form of tabulations and linear statistics (e.g., reporting of percentages and 



NISTNCSTAR 1-7, WTC Investigation 



Executive Summary 



average/means). The telephone interview results enabled a scientific projection of the population and 
distribution of occupants in WTC 1 and WTC 2, as well as causal modeling and multivariate regression 
analysis to explore fundamental egress issues such as sources of evacuation delay. 

The objective of the face-to-face interviews was to gather first-hand accounts and observations of the 
activities and events inside the buildings on the morning of September 11. This approach identified 
unknown information, aided in the evaluation of technical hypotheses, and explored motivations for 
occupant behaviors, while allowing for comparisons to the telephone interview data. There was no 
recording of the face-to-face interviews, other than random selections, with consent of respondents, for 
quality control purposes. A typical face-to-face interview averaged approximately two hours. The 
methodology for the face-to-face interviews was a synthesis of two established methodologies, designed 
to assist survivors in providing comprehensive and accurate accounts of their evacuation, given the 
latency between experience and interview. Some groups of occupants were specifically sought in order to 
explore targeted unknowns. These included occupants near the floors of impact, witnesses to fireballs, 
mobility-impaired occupants, floor wardens, building personnel with emergency response responsibilities, 
family members who spoke to an occupant after 8:46:30 a.m., and occupants from regions of the building 
not addressed by other groups in order to ensure adequate interview coverage for all areas of both towers. 

Six focus groups were conducted in order to elicit accurate group representations of specific events or 
themes and complement the findings of the telephone and face-to-face interviews. The focus groups and 
the corresponding objectives were: 

1. Occupants located near the floors of impact: to explore the extent of the building damage and 
how the damage influenced the evacuation process. 

2. Floor wardens: to explore the implementation of the floor warden procedures and the effect 
those actions had on the evacuation of the occupants on a floor and the evacuation of the floor 
warden. 

3. Mobility-impaired occupants: to explore the effect of a disability on the evacuation of the 
occupant and any other individuals who may have assisted or otherwise been affected by the 
evacuee. 

4. Persons with building responsibilities: to capture the unique perspective of custodians, 
security, maintenance, or other building staff. 

5. Randomly selected evacuees in WTC 1 : to further explore the variables from the causal 
modeling which best explained evacuation delay and normalized stairwell evacuation time, 
including environmental cues, floor, and activities. 

6. Randomly selected evacuees in WTC 2: to further explore variables used in the causal 
modeling that best explained evacuation delay, including environmental cues, floor, risk 
perception, and use of elevators. 

NIST documented the WTC egress system, including the location of the three primary stairwells, exit 
doors, core hallways, transfer corridors, wall construction, location and layout of the 100+ elevators in 
each tower, and emergency communication devices. The design of the egress system was compared to 



xxxii NIST NCSTAR 1-7, WTC Investigation 



Executive Summary 



building code requirements of the New York City Building Code, National Fire Protection 
Association 101 (Life Safety Code), and International Building Code. 

NIST documented the emergency procedures, both as they were planned to be carried out, as well as how 
they were actually implemented on September 11, 2001. The procedures included responsibilities for 
tenant safety through the floor warden system; pre-planned content of public address system 
announcements (which varied from public address system announcements made on September 11, 2001); 
responsibilities of the fire safety director, deputy fire safety director, building security, and supervisors of 
various contractors (including mechanical, vertical transportation, and electrical). Additionally, 
interaction among responding agencies such as the Port Authority of New York and New Jersey, the Port 
Authority Police Department, the New York City Police Department (NYPD), the New York City Fire 
Department, and contract security were documented. 

NIST estimates that there were 8,900 ± 750 people in WTC 1 at 8:46:30 a.m. on September 11, 2001. 
Similarly, NIST estimates that there were 8,540 ± 920 people inside WTC 2 at 8:46:30 a.m. New York 
City officially announced 2,749 fatalities at the WTC complex, including emergency responders, airplane 
passengers and crew (but not hijackers), and bystanders. NIST estimated that of the 17,400 ± 1,180 
occupants inside WTC 1 and WTC 2 at 8:46:30 a.m., 2,146 to 2,163 perished. No information could be 
found for 17 persons. More than twice as many occupants were killed in WTC 1 as WTC 2, largely due 
to the fact that occupants in WTC 2 used the 16 minutes between the attacks on WTC 1 and WTC 2 to 
begin evacuating, including the use of elevators by some occupants in WTC 2. 

The demographic characteristics of the evacuees was explored where the characteristics were relevant to 
the evacuation on September 11, 2001. Few differences in the characteristics of WTC 1 or WTC 2 were 
observed. Men outnumbered women roughly two to one. The average age was mid-forties. The mean 
length of employment at the WTC site was almost 6 years, while the median was 2 and 3 years for 
WTC 1 and 2, respectively. Sixteen percent of 2001 WTC evacuees were also present during the 1993 
bombing, although many other occupants were also knowledgeable about the 1993 evacuation. Two- 
thirds of the occupants had participated in at least one fire drill during the 12 months immediately prior to 
September 11, 2001. Eighteen percent did not recall whether they had participated in a fire drill during 
that time period and 18 percent reported that they did not participate in a fire drill during that time period. 

In WTC 1 , all three stairwells and the elevators were destroyed in the impact region, extending as low as 
floor 92. No occupant evacuated from above the 91st floor, although some survived until the building 
collapsed after 102 minutes. Helicopter rescue from the roof was considered by an NYPD aviation unit, 
but deemed not possible due to the heat and smoke from the building fire. Occupants of both towers 
delayed initiating their evacuation after WTC 1 was hit. In WTC 1 , the median time to initiate evacuation 
was 3 minutes for occupants from the ground floor to floor 76, and 5 minutes for occupants near the 
impact region (floors 77 to 91). Occupants observed various types of impact indicators throughout the 
building, including wall, partition, and ceiling damage and fire and smoke conditions. The most severe 
damage was observed near the impact region, fatally trapping some occupants. Announcements in 
WTC 1 were not heard by the occupants, despite repeated attempts from the lobby fire command station 
to order an evacuation. Damage to critical communications hardware prevented announcement 
transmission. Evacuation rates reached a peak, steady-state in approximately 5 minutes, and remained 
roughly constant until the collapse of WTC 2, when the rate in WTC 1 slowed to about one-fifth of the 



NIST NCSTAR 1-7, WTC Investigation xxxiii 



Executive Summary 



peak, steady-state. WTC 1 collapsed at 10:28:22 a.m., resulting in approximately 1,500 occupant deaths, 
107 of which were estimated to be below the 92nd floor. 

The evacuation of WTC 2 was markedly different from the evacuation of WTC 1 . There was a 1 6 minute 
period after WTC 1 was attacked, but before WTC 2 was attacked. During this time period, occupants 
were forced to decide whether to remain inside WTC 2, and if they decided to leave, they had to choose 
between using one of the three stairwells or using an elevator. Further complicating this decision process 
were multiple, conflicting announcements around 9:00 a.m., first instructing occupants to return to their 
offices, and then within one minute of impact, instructing them to begin an evacuation if conditions on 
their floor warranted that decision. Over 90 percent of WTC 2 survivors started to evacuate the building 
prior to its being attacked. Sixteen percent of the survivors used elevators to evacuate. Approximately 
75 percent of the occupants who were above the 78th floor (the lowest floor of impact) descended to at 
least below the impact region prior to the attack on WTC 2. Over 40 percent of the survivors had left 
WTC 2 prior to 9:02:59 a.m. After WTC 2 was attacked, at least 18 individuals used Stairwell A, located 
in the northwest corner and furthest from the impact damage, to descend below the 78th floor to evacuate 
the building. Additional public address announcements were made after the airplane strike on WTC 2, 
although occupants who survived generally did not hear those announcements. After the initial peak in 
evacuation rate, the rate reached a steady-state similar to the rate observed in WTC 1 until approximately 
20 minutes prior to collapse of WTC 2. The evacuation rate during the final 20 minutes dropped 
significantly, likely due to a decreased number of occupants remaining in the egress system below the 
78th floor. NIST analysis indicated only 1 1 occupants initially below the 78th floor were killed when 
WTC 2 collapsed at 9:58:59 a.m. Overall, NIST estimated that 630 occupants of WTC 2 perished. 

Using the statistical power of the telephone interview results, causal models were constructed to explain 
both evacuation initiation delay and average stairwell travel time per floor. The factors that best predicted 
evacuation initiation delay in WTC 1 were (1) which floor the respondent was on when WTC 1 was 
attacked, (2) whether occupants encountered environmental cues, and (3) seeking additional information 
(or milling) about the nature of the event. In WTC 2, the same process occurred as in WTC 1, except that 
perceived risk (sense of immediate danger) was a predictor of seeking additional information (along with 
floor and environmental cues). Analyses explored factors that affected time spent in the stairwells in 
WTC 1 exiting the building. The floor an occupant was on when WTC 1 was attacked (distance to safety) 
increased the probability of encountering an environmental cue (smoke, damage, fire, etc). Additionally, 
being on a higher floor predicted greater evacuation initiation delay times and encountering 
environmental cues, which predicted higher normalized stairwell travel time. Independently, interrupting 
evacuation for any reason increased the normalized stairwell travel time. 

Constraints or aids to the evacuation progress were documented. Building announcements were cited by 
many in WTC 2 as a constraint to their evacuation, principally due to the 9:00 a.m. announcement 
instructing occupants to return to their work spaces. Crowdedness in the stairwells, firefighter 
counterflow, lack of instructions and information, as well as injured or disabled evacuees in the stairwells 
were the most frequently reported obstacles to evacuation. The most commonly mentioned forms of aid 
were assistance from coworkers and emergency responders and the photoluminescent markings in 
stairwells. Six percent of survivors in WTC 1 and WTC 2 reported a mobility impairment which slowed 
their evacuation. Sometimes the evacuation speed of others in the immediate area slowed down occupant 
evacuation speed. Recent pre-existing injuries, medications, or medical treatments were the most 
commonly reported mobility impairments, while a small number used wheelchairs, were pregnant, or 



xxxiv NIST NCSTAR 1-7, WTC Investigation 



Executive Summary 



were elderly. A rest station for mobility-impaired occupants was established in WTC 1 somewhere 
between floors 12 and 20. Less than 10 minutes prior to the collapse of WTC 1, the occupants and 
helpers on the floor were ordered to evacuate, although it remains unclear whether all rest station 
residents survived. 

Minutes prior to the collapse of WTC 2, an NYPD Emergency Services Unit (ESU) officer radioed from a 
floor in the 20s to the outside that he was having trouble ascending the stairwell due to the large number 
of occupants descending (Interview 24 NYPD [NIST 2004]). While the origin of the occupants remains 
unknown, only 1 1 occupants who started evacuating below the impact region were known not to have 
survived. 

Multiple evacuation models were used to simulate different WTC tower evacuations, subject to a number 
of assumptions. The goal of the modeling was to frame an understanding of actual evacuation findings on 
September 11, 2001. Simulations demonstrated that a phased evacuation (also known as defend-in-place, 
whereupon occupants on the fire floor and the immediately surrounding floors descend to three floors 
below the fire floor) would have taken between 4 minutes to complete (without delays in evacuation 
initiation) and 11 minutes to complete (with evacuation initiation delays between and 10 minutes). 
Total evacuation of a tower assuming a full occupant load without visitors (19,800) would have required 
as few as 92 minutes to 1 12 minutes. With visitors (total population 25,500 people) total evacuation 
would have required as little as 1 14 minutes to 142 minutes. The ranges reflect two different model 
outputs, each assuming two different delay times (no delay and a 10 minute distribution of delay times). 
An evacuation simulation for 8,800 people (approximately the number present in each tower on 
September 11, 2001) in the absence of any damage to the building, would have required at least 
52 minutes to 71 minutes, depending on the model or the delay times. Finally, the EXODUS model was 
'calibrated' to approximate the gross evacuation rates observed in WTC 1 and WTC 2 on 
September 11, 2001. Once the model input necessary to approximate the observables was determined, 
additional occupants were added in order to estimate how many occupants might have been unable to 
evacuate on September 11, 2001 (given the damage to the building and observed delay times) if the 
buildings had had larger occupant loads. NIST estimated that approximately 14,000 occupants would 
have been unable to evacuate from WTC 1 and WTC 2 on September 11, 2001, had the starting building 
population been 19,800 in each building. 

E.2 REFERENCE 

NIST (National Institute of Standards and Technology). 2004. NIST WTC Emergency Responder 
Interview Data Set. Gaithersburg, MD. 



NISTNCSTAR 1-7, WTC Investigation 



Executive Summary 



This page intentionally left blank. 



xxxvi NIST NCSTAR 1-7, WTC Investigation 



Chapter 1 

Background and Introduction 



On the morning of September 11, 2001, the World Trade Center (WTC) in New York City was attacked 
by hijacked commercial airplanes. The collision with each tower (WTC 1 at 8:46:30 a.m. and WTC 2 at 
9:02:59 a.m.) produced significant structural damage. The impact generated a large, luminous external 
fireball that consumed a portion of the jet fuel, with the remaining fuel acting as an ignition source for the 
combustible material within each tower. At 9:58:59 a.m., 56 minutes after it was struck, WTC 2 
collapsed due to a combination of the aircraft impact damage and subsequent fire. WTC 1 stood until 
10:28:22 a.m. 

This report provides an analysis of the overall evacuation of WTC 1 and WTC 2. The two towers were 
nearly identical buildings in height, geometry, and architectural features. The evacuation processes in 
these two buildings displayed both distinct similarities and differences. This report also focuses on the 
behaviors of the occupants, actions of the building personnel and emergency responders (covered more 
completely in NIST NCSTAR 1-8 1 ), and the interactions among all three. This report documents the 
performance of the emergency egress system. 

This chapter begins with a discussion of significant egress events. It then reviews the design of the 
WTC egress system and emergency procedures, and outlines the National Institute of Standards and 
Technology (NIST) data collection methodology. The events of September 11, 2001, are detailed as they 
relate to the evacuation, including the moments prior to 8:46:30 a.m. Finally, an analysis of the events 
establishes key evacuation findings and conclusions. Note that individuals shown in photographs in this 
report may have been blurred to protect their identities. 

1.1 HISTORICAL INCIDENTS IN WHICH EGRESS WAS SIGNIFICANT 

Although the World Trade Center building collapses are arguably the most significant building events 
where building egress played a critical role, concern about the ability of occupants to escape from large 
buildings is hardly new. Indeed, many earlier lessons were based on analyses of high-rise fires in New 
York City. In 1911, the Triangle Shirtwaist Fire spread through the top three floors of a 10-story fire- 
resistant building in New York (Fire Engineering 1977). The fire started in a corner of the eighth floor of 
the building and quickly spread over the entire floor as well as the floors above by the windows, stairs, 
and elevator shafts. There were 145 fatalities in the fire, all but one from the ninth floor of the building. 
While many of the fatalities were located on the ninth floor, approximately 40 jumped from the building 
to the street below to escape the flames, and another 10 perished when an exterior fire escape collapsed. 
The fire was extinguished with hose lines from two standpipe risers in the stairwells and was under 
control within 18 min. The upper three floors were a complete loss. Significant issues identified from the 
fire investigation included the fact that there was limited access to the stairwells due to partially-blocked, 
non-fireproof doors that opened inward, as well as exterior cast-iron fire escapes which loosened from the 



This reference is to one of the companion documents from this Investigation. A list of these documents appears in the Preface 
to this report. 

NIST NCSTAR 1-7, WTC Investigation 1 



Chapter 1 



wall due to heat from the fire. Subsequent recommendations promoted fireproof egress stairways and 
automatic sprinklers for buildings taller than 18 m (60 ft). 

The Equitable Building fire in New York City in January 1912 is an early example of building collapse 
following fire (NFPA 1912). The Equitable Building was a group of five linked buildings, the tallest 
being 10 stories tall. Erected beginning in 1869, the buildings were constructed of so-called fireproof 
construction with wood floors on brick or tile arches supported by wrought-iron and steel I-beams resting 
on columns made mostly of cast iron. The fire originated in the basement of one of the buildings from a 
discarded match and spread quickly to a tile-enclosed shaft containing two elevators and eleven small 
dumbwaiters enclosed in wood. Within 45 minutes, the fire had spread throughout the upper floors of the 
buildings and downward through numerous unprotected floor openings. Except for a few areas, the 
building was completely gutted by the fire. Three separate sections of the building collapsed, with the 
largest collapse involving all of the floors down to the basement on one side of the building. Since the fire 
occurred before business hours, loss of life was limited to three employees on the upper floors and three 
additional deaths attributed to collapse of cast-iron columns. Firefighter loss was limited to a single 
fatality, as all personnel were ordered out of the building prior to the first collapse. Egress and firefighter 
access was through a single continuous stairway from the basement level to the top floor, deemed 
inadequate for escape in the subsequent investigative report. Recommendations included the need for 
protection of floor openings, corridor partitions, and structural metal work, and inclusion of sprinklers in 
all portions of office buildings where fire is most likely to occur. Two remote stairways enclosed in 
fireproof shafts with fire doors at each floor were deemed necessary. Additional stairways were 
recommended such that travel distance to a stairway was limited to 27 m (90 ft). 

In 1945, a U.S. Army Air Force B-25 crashed into the Empire State Building in New York City resulting 
in a significant fire on parts of the 78th and 79th floors from an estimated 3 m 3 (800 gal) of gasoline 
sprayed from the plane crash (Hayne 1945). The crash and resulting fire caused 14 deaths and 
approximately 25 injuries. The crash occurred on a Saturday morning when few building occupants were 
present, and much of the office space surrounding the crash site was unoccupied. Several occupants of the 
79th floor took refuge in a metal and glass partitioned office and were later rescued by the fire 
department. According to the investigation report, the stairwells remained tenable throughout the incident 
and provided fire department access and a safe means of egress for occupants of the upper floors not 
involved with the initial gasoline fire. Fire department access was accomplished via elevator to the 
65th floor and by stairwell the remaining 13 to 14 floors. The fire was extinguished approximately 
35 minutes after the first fire department notification. Building design, timing of the fire on a Saturday 
morning, and fire department response were credited with limiting the resulting damage and loss of life. 
Important issues related to building egress identified in the investigation report include (1) limiting use of 
elevators as a means of egress from upper floors, since the crash of one of the elevators to the sub- 
basement might create apprehension of the dependability of the remaining elevators; (2) an understanding 
that damage to stair shafts may be sufficient to prevent their use as a means of egress from the crash floor 
and floors above; and (3) a realization that fire resistive building construction does not preclude damage 
by fire involving building contents. 

In August 1970, a fire at the 50-story One New York Plaza building extensively damaged the 33rd and 
34th floors and spread significant smoke throughout the building (Powers 1970). The fire was first 
detected in the concealed ceiling space of the 33rd floor and spread to exposed polystyrene insulation in 
the south and west walls of the 33rd floor. The building was only partially occupied at the time of the 



NIST NCSTAR 1-7, WTC Investigation 



Background and Introduction 



fire, with some of the floors above the fire unoccupied. Occupants evacuated either by elevators or down 
the stairwells. Heavy smoke conditions were noted on many floors of the building. Two security guards 
and two firefighters died from the fire, and 30 injuries resulted. Fire department access was accomplished 
via elevators to the 30th floor and by stairs to the fire. The fire was controlled within 5 h. Reducing the 
fire load of building contents, the need for automatic sprinkler systems, and the protection of steel 
members by materials that cannot be readily removed or damaged were important issues identified from 
the investigation. New York City Local Law No. 5, Fire Safety in High-Rise Buildings, resulted in large 
part from a reaction to this and several other high rise-fires in New York at the time. Among other 
provisions, Local Law 5 requires building compartmentation, with an exception for sprinklered spaces 
(New York City 1973). 

In February 1972 and February 1974, major high-rise fires occurred in Brazil, causing more than 
200 casualties. In February 1972, a fire in the 31 -story Andraus Building in Sao Paulo resulted in 16 
fatalities and more than 375 injuries (Willey 1972). The fire developed on four floors of a department 
store and then spread up the exterior facade of the building, involving 28 floors of the building within 
25 minutes. The fire gutted most areas of the building and damaged structural supports. The department 
store occupied the lower seven floors above grade and was served by four open stairways and two 
elevators. The remainder of the building was of office occupancy with a single 1 m (39 in.) wide 
enclosed masonry spiral stairwell and five elevators. Door construction in the office stairwell was of 
hollow-core, wood, or metal construction. Combustible interior finish and exterior facade were credited 
for the rapid fire spread throughout all but the upper four floors of the building. It was reported that some 
people used elevators to egress the building, while others used the single stairwell. Once a stairwell door 
on the fifth floor failed, leaving the lower floors of the stairwell untenable, occupants fled toward the roof 
of the building. Approximately 300 people reached the roof level heliport, while another 200 became 
trapped in the stairwell. Rescue operations for those trapped in the stairwell included ladders from nearby 
buildings on the fifteenth and sixteenth floor. The use of areas of refuge by nearly 500 occupants was 
aided by stairway ventilation and wind velocity. 

An unfortunately similar fire two years later, which started on the 12th floor of the 25 story Joelma 
Building in Sao Paulo, and resulted in 179 deaths, 300 injuries, and total destruction of the building 
contents (Sharry 1974). Inability of helicopters to rescue occupants trapped on the roof of the building, 
inadequate means of egress from the building (a single 1.2 m (47 in.) unenclosed stairwell), lack of fire 
protection, and presence of combustible contents within the building were noted as significant in the fire. 
The majority of survivors of the fire made their escape through the use of the building's four elevators. 
While this method was not recommended due to the possibility that occupants may become trapped, the 
success of the evacuation was attributed to two factors: the use of elevator operators allowed the 
elevators to be operated in an express mode (stopping only at desired floors), and the elevator power 
supply was unaffected early in the fire. 

In June 1989, a fire occurred in a 10-story office building in Atlanta, Georgia (Isner 1990). The Peachtree 
25th building was an H-shaped building with two connected 10-story towers and a population of 
approximately 1,500 people. Each tower measured approximately 76 m by 20 m, with the connection 
measuring 21 m by 24 m. The fire began on the 6th floor of the south tower at approximately 10:30 a.m. 
The ignition of the fire was attributed to an electrician working on an electrical switchbox. While the 
worker was attempting to return power to a section of the floor by replacing a 200-ampere fuse, severe 
arcing occurred. The arcing had sufficient energy to melt metal and ignite the interior- finish materials in 



NISTNCSTAR 1-7, WTC Investigation 



Chapter 1 



the hallway. The electrician was severely injured and later died, although not as a direct result of the arc, 
which was estimated as having lasted 60 seconds or more. The fire growth rate was extremely high, and 
the fire spread was rapid. Multiple layers of wall covering promoted extraordinary fire spread rates, 
which was not an unfamiliar fire hazard to fire investigators (Bouchard 1982; Demers 1980). The wall 
coverings had completely burned out when the fire department arrived on the floor, only seven minutes 
after notification. The intense black smoke quickly trapped about 40 occupants on the floor of origin. 
Most occupants found a room and closed the door behind them, breaking out windows to vent incoming 
smoke and waited to be rescued. At some point, one woman jumped from a 6th floor window and 
sustained severe injuries. The fire department was not notified until an occupant of the building from a 
remote floor activated a manual pull-station at approximately 10:30 a.m. Several occupants of the fire 
floor were leaning out of a window in order to breathe when the fire department arrived on the scene at 
approximately 10:34 a.m. Fourteen occupants were rescued via ladder truck, and 14 people were rescued 
using the stairwells. In all, five people died because of this fire, the first multiple fatality high-rise office 
building fire in the United States in 17 years (Isner 1990). 

Several failure modes contributed significantly to the severity of the fire. There were no automatic 
sprinklers, which allowed the fire to spread. The electrician did not follow proper procedure when 
changing the fuse, resulting in the arc that ignited the wall linings and electrical equipment. The ignition 
source was so severe that a fire in the electrical room was inevitable, however. Multiple-layer, 
combustible interior- finishes also contributed to the rapid spread of the fire. 

In many instances, these significant egress events resulting from fires in buildings have shaped building 
codes requirements related to the egress system. Requirements for stairwell design, placement, and 
capacity all evolved as a result of significant past fire incidents. 

1.2 PREVIOUS FIRES AND EVACUATION INCIDENTS IN THE WORLD 

TRADE CENTER 

In February 1975, a fire in WTC 1 began on the 1 1th floor and ultimately spread from the 10th to the 
19th floor extending through telephone closets on each floor (Powers 1975). Although not important 
from an egress perspective, the fire provides an appropriate background for what occurred in the later 
terrorist attacks in 1993 and 2001. The fire was initially reported by manual alarm at 11:35 p.m. 
Automatic alarms from smoke detectors on the 11th floor through the 19th floor responded at about 1 min 
intervals after the manual alarm. It was believed that the fire originated in an executive office on the 
1 1th floor and spread to upper and lower floors through 0.30 m by 0.45 m (12 in. by 18 in.) openings in 
the floors of utility closets on each floor. Four steel floor trusses were distorted slightly. Approximately 
800 m 2 (9,000 ft 2 ) of the 11th floor was damaged, destroying about half of the contents and damaging the 
remaining contents in this area. Virtually all combustibles, including fire retardant-treated wood paneling 
on the telephone closet walls of the 10th and 12th floors, were destroyed. Limited quantities of 
combustible furnishings on the 12th and 13th floors limited the spread of fires from the telephone closets 
on these floors. Recommendations resulting from the fire included (1) provisions for automatic sprinklers 
in areas where highly combustible material or large accumulations of combustibles are present, (2) the 
installation of detectors in return air shafts on each floor to purge the return air and stop the supply of 
fresh air to the fire area, and (3) fire stopping of all openings in floors or walls as well as in any wiring 
installations. It was noted that sprayed fire-resistive material may not adhere properly to surfaces or may 
be dislodged as other building services are installed. 



NIST NCSTAR 1-7, WTC Investigation 



Background and Introduction 



On August 3, 1977, two Fuerzas Armadas de Liberacion Nacional (F.A.L.N.) terrorist bombs exploded in 
midtown Manhattan, killing one person and injuring seven others. When a specific threat against the 
World Trade Center was phoned into a local TV news station at 9:45 a.m. (Breasted 1977), both 
WTC towers were evacuated, although not until after 12 noon. An employee of Windows on the World at 
the time, described the situation: 

"We were all scared. I started to shake. The ride down seemed to take 
two hours. I'm part of a team that was trained for fire drills, but I have 
no idea of what do if there was a bomb. This was more frightening than 
a fire because we are all equipped for a fire." (Ivins 1977) 

An estimated 35,000 people were evacuated from WTC 1 and WTC 2, both of which reopened the same 
day, shortly after 3:00 p.m. (Ivins 1977). Overall, more than 100,000 people evacuated buildings in 
Manhattan that day (Breasted 1977). Many people, however, were reluctant to leave after having been 
docked wages after previous incidents for evacuating the building (Ivins 1977). 

At 12:18 p.m. on February 26, 1993, a terrorist attack resulted in an explosion in a sublevel parking 
garage in the World Trade Center complex, immediately killing six people (Isner and Klem 1993a; Isner 
and Klem 1993b) and causing an estimated $300 million damage. The explosion of at least 450 kg 
(1,000 lb) of explosive material caused extensive damage to several sublevels of the building and an 
intense fire that spread varying amounts of smoke in four of the seven buildings in the complex. Most of 
the complex's estimated 150,000 occupants evacuated the buildings as a result of the incident, including 
approximately 50,000 from the affected towers. 2 According to the NFPA Investigation, 1,042 people 
were injured in the incident, including 15 who received blast-related injuries. At the peak of the incident, 
the fire reached 16 alarms and involved more than 700 firefighters (approximately 45 percent of the New 
York City Fire Department's on-duty personnel) (Isner and Klem 1993a). As a comparison, on 
September 11, 2001, 22 alarms were called prior to the collapse of WTC 2, in addition to a 10-60 alarm 
(unique to special operations for large incidents) and a three alarm which staged additional units nearby. 
This resulted in the involvement of more than 1,000 firefighters being at the World Trade Center. 

The explosion significantly damaged floors, walls, and doorways in subgrade levels and forced large 
amounts of smoke well away from the immediate area. In one report, visibility was reduced to 0.3 m 
(1 ft) within about 1 min at the 44th floor of WTC 1, largely through the spread of smoke in elevator and 
stairwell shafts (Isner and Klem 1993b). Before beginning evacuation, many occupants experienced 
smoke on occupied floors and encountered even heavier smoke as they descended the buildings in the 
stairwells. Since the explosion disabled the emergency communication systems in the buildings, 
occupants responded to the event without the planned central guidance. Even without guidance, many 
occupants began evacuation early in the event. Egress was further complicated by a total loss of electrical 
power to emergency stairwell lighting within about 1 hour and 1 5 min. It was estimated that it took 
occupants from IV2 hours to 3 hours to exit the building from the upper floors of the towers. Fortunately, 
the scarcity of combustibles in the subgrade levels and dilution of the fire gases limited the toxic potency 
of the resulting smoke. Although most of the injuries were smoke related, no fatalities due to smoke 
inhalation were noted even with prolonged exposure to dense smoke. 



Text of undated presentation by Ted Stam, General Property Manager, World Trade Center. WTCI-619-P. 



NISTNCSTAR 1-7, WTC Investigation 



Chapter 1 



Fire crews were assigned responsibility for searching five floor subsectors. Since the elevators were not 
operational, firefighters climbed the stairwells. It took more than two hours for crews to climb to the 
100th floor. By 4:00 p.m., approximately 4 hours after the blast, all occupants had evacuated tenant 
floors. Some elevator cars, however, had stopped in elevator shafts, portions of which had no doors 
leading from that section of the shaft (such as express elevators). Locating and evacuating the trapped 
elevator occupants was such a high priority that, rather than wait for the stairwells to empty to send the 
elevator technicians up into the building by stairwells, the technicians were delivered to the roof by 
helicopter. Even with that measure, it took approximately 5 hours to locate and free a group of 
kindergarten students and several adults trapped in an elevator in WTC 2 (Isner and Klem 1993b). By 
8:00 p.m., approximately 8 hours after the explosion, the last occupants trapped in elevators were 
evacuated. 

1.3 SCOPE OF OCCUPANT BEHAVIOR, EGRESS, AND EMERGENCY 

COMMUNICATIONS PROJECT FOR THE NIST WORLD TRADE CENTER 
INVESTIGATION 

The purpose of this project was to determine the behavior and fate of occupants and responders - both 
those who survived and those who did not - by collecting and analyzing information on occupant 
behavior, human factors, egress, and emergency communications in WTC 1, WTC 2, and WTC 7, and 
evaluating the performance of the evacuation system on September 11, 2001. 

This project was divided into six tasks as follows: 

Task 1 — Gathered baseline information on the evacuation of the WTC buildings on September 11, 2001, 
through a comprehensive, systems-oriented, and interdisciplinary data collection effort focused on 
occupant behavior, human factors, egress, and emergency communications (including instructions given, 
interpretation of instructions, and response to instructions). This involved the collection of new data from 
people affected by the WTC attacks (e.g. building occupants, building operators, and emergency 
responders via direct accounts from survivors and families of victims), especially those who had to 
evacuate the buildings. Experts in human behavior and statistical sampling were retained to assist in 
developing a data acquisition strategy that considered various data collection methods, such as interviews 
and questionnaires. Inputs and suggestions were obtained from individuals with an interest in the data 
collection effort. Additionally, written accounts, transcripts of (emergency) communications, published 
accounts, and other sources of egress related information were obtained, in coordination with other data 
collection efforts for the overall investigation. 

Task 2 — Collected archival records from prior WTC evacuation incidents (e.g., 1975 fire, 1977 blackout, 
1980 bomb scare, 1990 power outage, and 1993 bombing) and practice evacuations, including oral history 
data from floor wardens and fire safety directors. These records were compared and contrasted with the 
September 11, 2001, evacuation. Changes made to the evacuation procedures following the earlier 
incidents and in recent years were evaluated in the context of the experience on September 11, 2001. 

Task 3 — Documented pre-event data for WTC Buildings 1, 2, and 7. This information included physical 
aspects of building egress components, such as stairs (width, number, location, vertical continuity), 
evacuation lighting, back-up power, elevators (number, operational before and after impact, role in 
evacuation), and active fire protection systems (sprinklers, manual suppression, fire alarms, smoke 



NIST NCSTAR 1-7, WTC Investigation 



Background and Introduction 



control). Building plans, emergency plans, type and frequency of evacuation drills, occupancy level and 
distribution on the morning of September 1 1th, and communications also constituted pre-event data. This 
information provided a baseline for evaluating the performance of the egress system. 

Task 4 — Stored the information collected in task 1 in a database. Additionally, information from third- 
party sources, such as published media accounts, were assembled and analyzed in the database. 

Task 5 — Analyzed the data to study the movement of people during the evacuations, decision- making and 
situational awareness, and issues concerning persons with disabilities. A timeline of the evacuation was 
developed using the results of these analyses together with other data sources. This timeline was 
compared with the timeline of the structural response, data on the development of the interior conditions 
(fire and smoke), as well as information on the activation of the active fire protection systems. The 
observed evacuation data was compared with results obtained using alternate egress models to better 
understand occupant behavior and identify needed improvements to existing egress models. In addition, 
the evacuation experience was compared with previous evacuation incidents in these buildings. The 
results were reviewed in the context of occupant protection practices for tall buildings, including the 
consideration of total evacuation and phased evacuation strategies. 

Task 6 — Report preparation. The results of this project were synthesized into this report to describe the 
occupant behavior, egress, and emergency communications in WTC 1 and WTC 2, and the performance 
of the evacuation system. 

1.4 REPORT ORGANIZATION 

This report investigates the occupant behavior, egress, and emergency communications at the World 
Trade Center on September 11, 2001. 

This chapter explores historical fire incidents where egress played a significant role, as well as previous 
significant fire or evacuation incidents at the World Trade Center complex. It also describes the scope of 
the overall project. 

Chapter 2 describes the design of the World Trade Center egress system, including the stairwells, 
elevators and emergency communication systems. Emergency procedures, including the roles of building 
managers is described. Finally, changes to the egress system as a result of the 1993 bombing are detailed. 

Chapter 3 documents the overall technical approach of the project, including discussion of the collection 
and analysis of first-person accounts (face-to-face, telephone, and focus group interviews), collection and 
analysis of published media accounts, and collection and analysis of other relevant data, including audio, 
video, photographic, and design records. 

Chapters 4 through 9 chronologically detail the overall progression of the evacuation of WTC 1 and 
WTC 2, including occupant activities, observations, and reactions. 

Chapter 4 enumerates the occupants of WTC 1 and WTC 2, describes their basic characteristics as it 
relates to evacuation, and discusses the emergency preparedness of the occupants prior to the attacks. 



NISTNCSTAR 1-7, WTC Investigation 



Chapter 1 



Chapter 5 documents the occupants' awareness of and reaction to the impact of the first airplane with 
WTC 1, as well as observations of local damage and phenomena. 

Chapter 6 describes the period of time from immediately after WTC 1 was attacked until just prior to the 
attack on WTC 2. The overall evacuation rate, actions of the building managers, and occupant activities 
and behaviors are discussed. 

Chapter 7, paralleling Chapter 5, documents the impact of the second airplane with WTC 2, including 
awareness and reaction on the part of the occupants. 

Chapter 8 tracks the progress of the evacuation and overall emergency response in both towers until 
immediately prior to the collapse of WTC 2. 

Chapter 9 examines the collapse of WTC 1 and WTC 2, including an analysis of where the occupants 
likely were as each building collapsed. 

Chapter 10 discusses the important egress issues raised by the events of September 11, 2001, at the World 
Trade Center. Included are causal models, summary statistics on the overall building evacuation rates, 
egress modeling, and in-depth analysis of specific issues that affected the evacuation, including the role of 
alarms, announcements, mobility impairments, emergency responders, authorities, information flow, 
activities, evacuation experience, and constraints/aids to evacuation. 

Chapter 1 1 summarizes the key findings of this report and highlights the most important findings to 
consider in response to the evacuation of WTC 1 and WTC 2. 



NIST NCSTAR 1-7, WTC Investigation 



Chapter 2 

Design of the World Trade Center Egress System 



The provision of access to and egress from buildings under emergency conditions relies on four primary 
components: stairwells, elevators, communication systems, and emergency responders (broadly defined to 
include the City of New York Fire Department (FDNY), New York City Police Department (NYPD), 
Port Authority Police Department (PAPD), Port Authority personnel, building security, fire safety 
directors, floor wardens, and other individuals with formal response responsibilities). These are 
subsequently grouped into building systems and the human component. 

2.1 OVERALL BUILDING DESCRIPTION 

By 2001, the World Trade Center (WTC) complex had become an integral part of Manhattan. It was 
composed of seven buildings (here referred to as WTC 1 through WTC 7) on a 1 6 acre site, located near 
the southwest tip of the island, shown in Figure 2-1. Whether viewed from close up, from the Statue of 
Liberty across the Upper Bay or from an airplane descending to LaGuardia Airport, the WTC towers were 
a sight to behold. WTC 1 (often referred to as the North Tower) and WTC 2 (often referred to as the 
South Tower), were each 110 stories high, dwarfing the other skyscrapers in lower Manhattan and 
seemingly extending to all Manhattan the definition of "tall" set by midtown's Empire State Building. 
Groundbreaking for the towers was in 1966, while construction began in 1968. WTC 1 was first occupied 
in 1970; WTC 2 in 1972. 

Additionally, there was a six-story subterranean structure, largely below the WTC Plaza with connections 
to WTC 1, 2, 3, 4, 5, and 6, which included a shopping mall and the WTC PATH station. This was 
surrounded by a 3 ft (0.9 m) thick concrete wall that extended from ground level down 70 ft (21 m) to 
bedrock. Holding back the waters of the Hudson River, this wall had enabled rapid excavation for the 
foundation and served to keep the groundwater from flooding the underground levels. Commuter trains 
brought tens of thousands of workers and visitors to Manhattan from Brooklyn and New Jersey into the 
WTC station. A series of escalators and elevators took the WTC employees directly to an underground 
shopping mall and to the Concourse Level of both towers. 

WTC 3 (Marriott Hotel) was 22 stories. WTC 4 (South Plaza Building) and WTC 5 (North Plaza 
Building) were both 9-story office buildings. WTC 6 (U.S. Customs House) was an 8-story office 
building. These six buildings were built around a 5 acre plaza, named for Austin J. Tobin, and the 
centerpiece of which was a large globe art object. WTC 7, located north of the other six WTC buildings 
and separated by Vesey Street, was a 47-story office building. WTC 7 was completed in 1987 and was 
operated by Silverstein Properties, Inc., as an air rights building. 



NISTNCSTAR 1-7, WTC Investigation 



Chapter 2 



Plaza Level 




WTC 3 WTC 2 




WTC4 



Bridge to 

130 Liberty Street 



Liberty Street 



Figure 2-1. WTC site plan. 



2.1.1 



Description of the Towers 



WTC 1 and WTC 2 each consisted of 1 10 stories above the Concourse Level (or 109 stories above the 
plaza / Mezzanine Level) structure. There were also six basement levels below the Concourse Level. 
Although the towers were similar, they were not identical. The height of WTC 1 at the roof level was 
1,368 ft (418 m) above the Concourse Level (6 ft taller than WTC 2), and WTC 1 additionally supported 
a 360 ft (1 10 m) tall antenna on the roof for television and radio transmission. Each tower had a square 
plan with the side dimension of 207 ft 2 in. (63.2 m). The corners of the tower were chamfered 6 ft 1 1 in. 
(2.1 m). Each tower had a core service area of approximately 135 ft x 87 ft (41 m x 27 m), although the 
core space changed on tenant spaces throughout the towers. A typical architectural floor plan in the tower 
is shown in Figure 2-2. As can be seen in this figure, placing all service systems within the core provided 
column-free floor space of roughly 31,000 sq ft (2,900 m 2 ) per floor outside the core. The long axis of the 



10 



NIST NCSTAR 1-7, WTC Investigation 



Design of the World Trade Center Egress System 



core in WTC 1 was oriented in the east-west direction while the long axis of the core in WTC 2 was 
oriented in the north-south direct. 




Figure 2-2. Typical WTC tower architectural floor plan. 

The superb vistas from the top of such buildings virtually demanded public space from which to view 
them, and the Port Authority responded. The 107th floor of WTC 1 housed a gourmet restaurant and bar 
with views of the Hudson River and New Jersey to the west, the skyscrapers of midtown Manhattan to the 
north, the East River and Queens to the east, the Statue of Liberty to the southwest, and the Atlantic 
Ocean to the south. Similar views could be seen from observation decks on the 107th floor and the roof 
of WTC 2. 



NISTNCSTAR 1-7, WTC Investigation 



11 



Chapter 2 



Table 2-1 shows the use of the floors, which was similar but not identical in the two towers: 

Table 2-1. Use of floors in the WTC towers 



Floor(s) 


WTC1 


WTC 2 


Roof 


Antenna space and window washing 
equipment 


Outdoor observation deck and window 
washing equipment 


110 


Television studios 


Mechanical equipment 


108, 109 


Mechanical equipment 


Mechanical equipment 


107 


Windows on the World 


Indoor observation deck 


106 


Catering 


Tenant space 


79 through 105 


Tenant space 


Tenant space 


78 


Skylobby, tenant space 


Skylobby, tenant space 


77 


Tenant space 


Tenant space 


75,76 


Mechanical equipment 


Mechanical equipment 


45 through 74 


Tenant space 


Tenant space 


44 


Skylobby, kitchen, tenant space 


Skylobby, tenant space 


43 


Cafeteria 


Tenant Cafeteria 


41,42 


Mechanical equipment 


Mechanical equipment 


9 through 40 


Tenant space 


Tenant space 


7,8 


Mechanical floors 


Mechanical floors 


Concourse through 6 


6-story lobby 


6-story lobby 



The Port Authority had managed the operation of the two towers since their opening three decades earlier. 
Silverstein Properties acquired a 99-year lease on the towers in July 2001. 

At the beginning of the workday, many of the roughly 40,000 people who worked in the towers and 
visited to tour or to conduct business emerged from PATH trains in the massive subterranean station. 
They would take escalators and elevators to a large shopping concourse. Walking a few hundred feet led 
occupants to the spacious, 6-story-high lobby on the Concourse Level where they would cross paths with 
those who arrived on foot or by bus and cab. Figure 2-3 shows the layout of the shopping mall, located 
underneath the WTC plaza. Figure 2-4 shows the lobby configuration for WTC 1. Figure 2-5 shows the 
layout of the WTC 2 lobby. The WTC 1 and WTC 2 lobbies were at the same level as the underground 
shopping mall, often collectively referred to as the Concourse Level. The WTC outdoor plaza and the 
WTC 1 and WTC 2 Mezzanine were one story higher than the Concourse Level, often referred to as 
either the Mezzanine or plaza level. 



12 



NIST NCSTAR 1-7, WTC Investigation 



Design of the World Trade Center Egress System 



Concourse (Mall) Area 



Vesey Street 




Li berty Street 

Figure 2-3. Shopping mall layout underneath WTC plaza. 



NISTNCSTAR 1-7, WTC Investigation 



13 



Chapter 2 




Concourse 



Figure 2-4. WTC 1 lobby (concourse) level. 



14 



NIST NCSTAR 1-7, WTC Investigation 



Design of the World Trade Center Egress System 



Concourse 



Escalators 



Down 




Figure 2-5. WTC 2 lobby (concourse) level. 



NISTNCSTAR 1-7, WTC Investigation 



15 



Chapter 2 



Getting tens of thousands of people from the Concourse to their offices was no small task. This was 
accomplished by a then-novel array of 106 express and local elevators located within the building core (as 
shown later in Figure 2-14). Section 2.2.3 discusses the elevators system in WTC 1 and WTC 2. 

Also within the core were three sets of stairs that extended the full height of the tower. Section 2.2.2 
discusses the stairwells in each tower. However, upon entering a stairwell at an upper floor, one did not 
find a continuously descending staircase leading to the lobby. Principally at the mechanical floors, there 
were enclosed horizontal corridors that led around the massive elevator hardware. These corridors ranged 
in length from about 1 ft to about 1 00 ft. After traversing each of these, the pedestrians would resume 
their descent. 

Upon exiting the elevators (or stairs, for those who chose the more strenuous route), one was faced with a 
view typical of high-rise buildings. Surrounding the rectangular core corridor was a mixture of blank 
walls, door entries to firms, and glass-front reception areas. Above was a standard drop ceiling. 

Many of the floors had but a single tenant, and some of these tenants occupied multiple floors. By 2001, 
most of these companies, which had moved in since the installation of automatic sprinklers, had taken 
advantage of Yamasaki's design concept of a vast space that was virtually obstruction- free. The open 
landscaping included as many as 200 or more individual workstations, often clustered in groups of six or 
eight (Figure 2-6). Trading floors had arrays of long tables with multiple computer screens (Figure 2-7). 
Some of these floors had a few executive offices in the corners and along the perimeter. Many also had 
walled conference rooms. It was common for the multiple-floor tenants to have installed convenience 
stairs internal to their space. 

Other floors were subdivided to accommodate as many as 20 firms. Some of the smaller firms occupied 
space in the core area, reclaimed as elevator shaft space from local elevators was phased out throughout 
the towers. 

With so many workers and visitors in the buildings, there needed to be food available. The Port 
Authority maintained a cafeteria on the 43rd floor of WTC 1. A number of the companies maintained 
kitchen areas where catered food was brought in daily, making it unnecessary for their staff even to leave 
their floor for lunch. The underground Concourse Level mall also provided may options for eating. In 
addition, there were hundreds of restrooms, in both the tenant and the core spaces. 



1 6 NIST NCSTAR 1-7, WTC Investigation 



Design of the World Trade Center Egress System 




Figure 2-6. Typical WTC tenant 
spaces. 



Source: Photos courtesy of The Port Authority of 
New York and New Jersey. 





Figure 2-7. A WTC 4 trading floor. 



Source: Photo courtesy of The Port Authority of 
New York and New Jersey. 



NISTNCSTAR 1-7, WTC Investigation 



17 



Chapter 2 

2.2 BUILDING SYSTEMS 

2.2.1 Egress Calculations 

Determining the Number and Width of Stairwells 

In 1965, architects and engineers designing the World Trade Center towers were faced with an impending 
change to the NYC Building Code. The draft building code language had a significant impact on the 
design of emergency egress systems. In 1965, the Port Authority directed its designers to adopt the draft 
version of the new code for their final designs. Some of the advantages of the new draft code were noted 
to be the following (Levy 1965): 

• Fire towers 3 could be eliminated; 

• Provisions for exit stairs were more "lenient;" and 

• Criteria for partition weights were more "realistic." 

It was not certain whether all the changes being proposed to the 1938 code would be incorporated into the 
final version of the new code. Thus, in 1966, the Chief Engineer of the Port Authority suggested that the 
"architect/engineers prepare a listing of the elements of the design which do not conform to old code 
requirements, but are acceptable under the new. With this list in hand, we could initiate discussions, at 
top level in the Building Department, to see if we can secure agreement to go along with our design" 
(Kyle 1966). 

A one-page document, dated "2/15/67", with the initials "CKP" listed the following items: 4 

• Fire tower corridors [sic] eliminated. 

• Number of stairs reduced from 6 to 3. (Old plans had 5 stairs at 3'-8" and 1 stair at 4'-8" for a 
total population of 390. 5 New plans have 2 stairs at 3'-8" and 1 stair at 4'-8" allowing a 
population of 390.) 

• The size of doors leading to the stairs are [sic] changed from 3'-8" to 3'-0". 

• All stairs exit through a lobby. Old plans had fire tower stair exiting through a fire enclosed 
corridor. 

• Shaft walls are changed from a 3 h rating to a 2 h rating. 

• Corridors are limited to a 100 ft dead end and with a 2 h rating. 



A fire tower is an exterior stairwell of incombustible construction terminating at grade level designed to ensure that smoke 
conditions from an interior fire do not contaminate the fire tower. The fire tower was provided for firefighter ingress and did 
not count as a required stairwell for occupant egress. (NYC Building Code 1938) 
See appendix of NCSTAR 1-1 for a reproduction of this memo. 
5 The 1938 NYC Building Code allowed 30 person per unit of exit width, while the 1968 NYC Building Code allowed 
60 persons per unit of exit width, effectively halving the egress capacity of new construction. Population calculations are 
per floor. 

18 NIST NCSTAR 1 - 7, WTC Investigation 



Design of the World Trade Center Egress System 



• Additional (word(s) missing) changed from 20 pounds per ft 2 to 6 pounds per ft 2 (based on 
partition weight of 50 pounds to 100 pounds per linear foot). 

Apparently, this list represented elements of the WTC design that would not have satisfied the 1938 code, 
but did satisfy the then-current draft version of the new code. Ultimately, WTC 1 and WTC 2 were 
designed with three stairwells, two 3 '-8" (44 in.) wide and one 4'-8" (56 in.) wide, as discussed below. 

A unit of exit width in the 1968 NYC Building Code was (and continues to be) 22 in. (0.56 m). The NYC 
Building Code table specifying exit and access requirements (Table 6-1) required that for a business 
occupancy, the stairs would accommodate 60 persons per unit of exit width. As the WTC 1 and WTC 2 
tenant floor design occupancy load was 365 persons per floor 6 (Solomon 1968), this required 6.5 units of 
exit width. Twelve in. (0.3 m) was the minimum half-width acceptable in the code, therefore, three 
stairwells (two with two units of exit width (44 in. [1.1 m]) and one with 2.5 units (56 in. [1.4 m])) 
satisfied the minimum requirements of the 1968 NYC Building Code. Table 2-2 shows the location of 
the stairwells, core perimeter, and transfer hallways for occupied floors in WTC 1 and WTC 2. 

Egress Provisions from Windows on the World 

The 106th and 107th floors of WTC 1 (North Tower) contained the Windows on the World complex, 
consisting of the Windows on the World restaurant, the Greatest Bar on Earth, numerous banquet and 
function rooms, kitchens and support areas, and management offices for the dining complex. While the 
configuration of the space may have changed over the life of the building, these functions were all present 
from the time Windows on the World first opened in April 1976. 7 

Restaurants, bars, and function rooms are classified in building codes as assembly use, which carries a 
significant increase in occupant load and consequent provisions for egress. The design occupant load for 
such assembly space is 1 5 ft 2 per occupant as opposed to the 1 00 ft 2 per occupant for the office use space 
in most of the rest of the buildings. Thus, while the design number of occupants on an office floor was 
365 to 390 (depending on the calculation method), the design number of occupants for these floors was 
over 1,000 each (the exact number depends on the area of kitchens, dishwashing, and office space on the 
floor, all of which is at 100 ft 2 per occupant). 

Locating assembly space high in a building poses particular challenges to egress design because the 
capacity of an egress component is not permitted to be decreased in the direction of travel. Thus, where 
more or wider stairs are provided to meet capacity requirements these must be continued all the way 
down through the building which affects space utilization for the entire structure. 



A January 25, 1968 memo from J. Solomon (Emory Roth and Sons) to M. Levy (PANYA) subsequent to a NYC Building 
Department plan review, documents that the "largest floor area is about 36,500 ft on the 106th floor. At one person per 
100 ft there will be 365 persons per floor, well within the permissible maximum" of 390 persons based upon stairwell 
capacity. WTCI-477-P Note that this calculation did not account for the use of the 106 th floor as an assembly space. 
7 PANYNJ response to formal NIST question, March 25, 2005. 

NIST NCSTAR 1-7, WTC Investigation 1 9 



Chapter 2 



Table 2-2. Plan view of stairwells in WTC 1 and WTC 2. 



.oi^ 1 "" 



1,111111"' 



,#«> 



• ^\ 




WTC 2 



WTC1 



I-- 


#l ^<^ N 




-.'«" 




© 

WTC 2 


V"^ 


WTC1 



Lobby (Concourse) Level Layout 
Only Stairwell B Serviced the Concourse Level 



MST 


X 


>v 


\ 


& 








© 

WTC 2 


\ 





f0*' 


© 

wrci 



Mezzanine (Plaza) Level Layout 
Only Stairwell B Serviced the Concourse Level 




WTC 2 



WTC1 



Floors 9-19 
Stairwells A was East in WTC 1, North in WTC 2. 




WTC 2 



WTC1 



Floors 20 - 26 

Core space previously used by the elevators in the northeast 

quadrant became leasable tenant space. 



tf## 



„\»»' 



,«i»' 



WTC 2 



Ull"' 



,1111" 



mill'"' 



,,,,,,1111 % 






"•«, 



<± & 




WTC1 



Floors 27 - 34 



Floors 35 - 41 



20 



NIST NCSTAR 1-7, WTC Investigation 



Design of the World Trade Center Egress System 



1-."— 




x 


\ 


t 




^ 


© 

WTC2 


\ 


10 


© 

WTCl 







)|||fl "Hi, 




C^ 


•O 




\ 


^ 




^ 


© 


V 


^ 


© 


WTC2 


n 


W 


WTC1 







Floor 42 




Stairs A and C transferred outside the core. 


^ 


*** 


^.--v 

- ^ 


\ 


S 


t u 


^ 


© 


\ 


tfH0^ 


© 


WTC2 


II 


w^ 


WTCl 



Floors 43, 45 - 47 
There was an escalator connecting floors 43-44 (skylobby). 



»!! 



,,11111'""" 




WTC2 



WTCl 



Floor 44 (Skylobby) 



Floor 48 
Stairs A and C transferred back inside the core. 



#""" "X, 



WTC2 



,^«""" /'X 




WTCl 







WTC2 



WTCl 



Floors 49 - 54 



Floors 55 - 56 



NISTNCSTAR 1-7, WTC Investigation 



21 



Chapter 2 



^ 




WTC2 



WTC1 




Floors 57 - 75 
There was a slight change in Stairs A, C between floors 66-68. 



Floor 76 

Stairs A, B, and C transferred, with Stairwells A and C moved 

outside the core. 





S \0^ 


© 

WTC1 



■*■ 






<<■ 




> 


WTC2 


* 


© 

WTC1 



Floors 77, 79, 80, 81 
There was an escalator connecting floors 77-78 (skylobby). 



Floor 78 




WTC2 







WTC2 



WTC1 



Floor 82 
Stairs A and C transferred back inside the core. 



Floors 83 - 95 



22 



NIST NCSTAR 1-7, WTC Investigation 



Design of the World Trade Center Egress System 



,„,,..-" 

r o% ^ 


© 

WTC2 


\X 


WTCl 




Floors 96 - 102 



ram 










& 


\ 


V 


<s> 


<> 


X 


\ 






V 


© 


^kama 


,;:::;:# 


© 


WTC2 


w 


F^ 


WTCl 





Floors 103 - 104 




x 


© N 

WTC2 


V^ 


© 

WTCl 



Floors 105 - 106 



Floor 107 



The document record contains a letter dated January 27, 1995, from Eugene Fasullo (PANYNJ) to 
Richard Visconti (Deputy Commissioner, NYC Department of Buildings [DOB]) confirming the results 
of a meeting on December 6, 1994, at which they reached agreement on a plan to address egress 
requirements from the 106th and 107th floors (Fasullo 1995). The details of the agreed solution are 
summarized below. The Deputy Commissioner, DOB, signed the letter to show concurrence with the 
agreed solution. 8 

It remains unclear what conditions existed from the date Windows on the World first opened to the time 
the agreed solution was implemented in 1995. The dates suggest that the issue was identified as a result 
of the Memoranda of Understanding (MOU) between PANYNJ and the NYC DOB and FDNY executed 
in 1993, in response to the bombing. A Windows on the World refurbishment after the 1993 bombing 
included these egress system changes. 9 

The basis for the agreed solution was to divide each floor into three areas of refuge (consistent with 
Section 27-372 [NYC Building Code]) to provide additional capacity to the existing stairs in accordance 



Fasullo, E., PANYNJ, to R. Visconti, NYC Department of Buildings, "Variance Granted by Memorandum of Understanding 
with Buildings Department, Windows on the World," January 27, 1995. 
PANYNJ response to formal NIST question, March 25, 2005. 



NISTNCSTAR 1-7, WTC Investigation 



23 



Chapter 2 



with Section 27-367 (NYC Building Code). These identical provisions existed in the version of the 
1968 NYC Building Code in effect when the buildings were built as sections C26-604.5 and C26-603.3, 
respectively (the NYC Building Code was renumbered) (NYCBC 1968). 

These code provisions allow for a doubling of allowed stair capacity when one area of refuge is provided 
on a floor and tripling the stair capacity for two or more areas of refuge on a floor. These areas of refuge 
must be separated by 2 h construction, be large enough for the expected occupant load at 3 ft 2 per 
occupant, each contain at least one stair, and have access to at least one elevator (above the 1 1th floor). 
Since three, distinct areas of refuge were provided on each floor, tripling of the capacity of each of the 
three stairs resulted in a maximum permitted occupant load of 1,170 people per floor (6.5 units of egress x 
60 persons per unit x 3). 

Attached to (and referenced in) the letter were two plans entitled "106th Floor Egress Plan" and "107th 
Floor Egress Plan" (shown in Figure 2-8 and Figure 2-9, respectively) that detailed the arrangement,. 
The 2 h separation walls snaked across the floors and were not aligned on the two floors. Some areas that 
needed to remain open to free passage were protected with Won doors (accordion doors that are fire rated 
and are closed automatically on activation of the fire alarm system). Details of the egress system design 
calculations and corresponding NYC Building Code requirements were included on the plans to 
demonstrate they met code requirements. 

By comparison, current model building codes, including the ICC 2003 International Building Code (IBC) 
and National Fire Protection Association (NFPA) 5000, both permit a doubling (but not tripling except in 
IBC Type 1-2 and 1-3 institutional uses) of the stair capacity for the provision of a horizontal exit on a 
floor. The horizontal exit must consist of a 2 h fire rated separation, contain at least one stair on each 
side, and have sufficient space for the expected occupant load at 3 ft 2 per person. A horizontal exit must 
be continuous down through the building to grade 10 (NFPA 1 1.2.4.3.1 and IBC 1021.2), unless the floor 
assemblies are at least 2 h with no unprotected openings. 

The solution to the egress problem was to provide a protected space in which occupants could wait to 
enter stairs that did not have adequate capacity for the numbers of people. Since the attacks took place in 
the morning (a non-peak time), NIST estimated that there were 188 occupants trapped in the Windows on 
the World floors. If the attacks had occurred when the facility was loaded near its capacity, as many as 
2,000 occupants could have lost their lives on those two floors alone, since there were no survivors above 
the impact floors of WTC 1 . 



In other words, the stairway may not contain unprotected openings (such as opening out to a floor) until the occupant exits the 
building. 

24 NIST NCSTAR 1 - 7, WTC Investigation 



Design of the World Trade Center Egress System 



III 

Be 

2: 




:.> .-rrnr.'-T J :-.- y, f. ,,-- ■ 
i-i t' 'it r* »i "• i 

l'jJ ,e JL&.-J j, |r ..^n- r 

' '•* r ' r is & ~ ' ■ ' i i * * : * ■ 

■ fij 4 ^ >! *tj ■!■ !l 

. ^_-. ,■-'.'. '^--^-.., .-.----■- _-w_.. 
„ . . ■ v i j j : j : . . 




I ■ 



llffiM 




Hi , III V^tl 
nil tt Jii -4 



- 



4 S 






™ tn 



ii Ml 

UJ o o 

a: = e 

= m- 
in irt ll 
<N "— £> 
II » U 

is S m 
woo 



ljj n q 

Q O S? 



@: 



CV ^ .& 
It ii "o 
^ li (0 

's 8 m 
to o o 



ft °- 
Ul is 

""- CM 
" II 

3 2^ 

go e 



I J £ 

U) o o 



tO 



S.S, 



■SS 



IMS 

_ OL "- m 
h^ C SC TT 
[^ i^ rt> a 



@Sg s 






■g in 






o 


c 

CO 


"o 

c 
o 





Q_ 


w 


£ 


s_ 


b 


+* 


o 


Q. 


c 


o 


-£= 


o 


Li. 


g 


5 

o 
■o 


(D 

O 


O 

cu 
o 

o 

L. 

Q. 


c 


"*~~ 


cr 






I* 5= b 





(O 




o 


r- 


rH 


U) 


^_ 


^ 


o 


>s 


o 


-£-! 


4— 






!_ 
<1> 


c 


F 


o 


CD 


+-> 


r 


CC 


cc 


-J 




o 


LU 


CS 




o 


<1) 


en 


0) 


en 


(1) 


u 


■-J 






TO 
CD 


^: 


T3 


"O 


i- 


!_ 


O 


o 


(1) 


> 


-C 


^ 


+J 


CD 


r 


^ 


o 






O 


en 


^ 


§ 


o 

r" 


o 

T3 


Z3 


C 



2 8 

cu at 



oo 

i 

CM 

CD 

;_ 
3 



0) 

o 

3 
O 
CO 



NISTNCSTAR 1-7, WTC Investigation 



25 



Chapter 2 




f 4~ 



CD 



LL < 






.6 I 




£ * 













IF 






r_j 


















O- 










s 


CO 








D- 

E>5 


O, 


m S 

*- Si 




cmj 

c 

£ 

u 
ai 
to 


A3 

cz 






& 

^ 
£ 


>l =) 


II 3 


11 II 

BE! 




D_ 








^ UJ <£ 


- } m £ 




Oi 


M 

■: 

1 




en 






g§ 1 


&g 




o 


LO 

C 
n 


.5 


to 3 

0)g 


ID (jgQ. 


tO @CL 


©*" 


a. 


en 


n 


L'- 


©=■8 


@ S R 


3 <B> 


35 




CM 


£ 


O 


g* 


= UJ — 




LU 3 






CD 


--- 






LU in ii 


LU lC. II 


in LU 


II 


m 


■ 




m 


CM *~~ ^* 


CM ■"- £? 


cm CM 


ft 1 








11 


ii II 'O 
■5^3. 


II H tj 

■5 ° ^ 


ii 11 




>- 






s 


* 


5 o t 


9 o to 




p.-. 




^3 

c 

< 






8. 
to 

1 


CO Q O 
6 


d 


£j0 a 
l*i 
d 


(J 


TJ 




Zi 

Q 



3. 



T3 "■ 

™§ 

— c 
a) £ 

5 e 
e s 



[1 

: 1 



5£ 



I ij 



< 





fl 






i 


1 










3 


3 




^ JS K 






M 


ff 




a? g> 






£ 


dE 








e 


€ 










n 


ED 


in 


S s 




s 


s 


SR 


ia 






—i 


01 


J u. 


■CM 


■ri . L ; 




3 


S 


a s 




ii 




1 




«! 


4) 


SM 


^ 


ft 




n%z 


Ql 


S.HJ 


C3 




UJ^ 


S.S I 


O 






r/l 


" 3) 


U5 " 5) 


a 

en 




LJ 
fl 


o 

F 
- 


Q_ :ii 


US 


y 


■i^ 


1TJ 

a> 


< 
Q- 




< O A3 






m 


I 


^ ni 




s 


cm £ 


,-' 


U- P 


rt u_ P 














E 


^ 










Q 




O 




o 




^ 




K 




2 




2^ 




>n 








jB 




JU 




3S 


E 




F 




E 


V 




q 




9 


*o 


1 




I s 




i 


s 
e 


"5 




■5 




"5 


S 




s 




S 


z 


o_ 




D_ 




fl_ 



CD D_ 

*■ O 

c o 
to - 

^^ 

° ?; 





r» 




o 


H 


rH 


(/) 


^_ 


^ 


o 


>> 


o 


-£J 


4— 






!_ 
(1) 


c 


1= 


o 


QJ 




O 
C 


CC 


CC 


-J 




o 


LJJ 


CC 




o 


(1) 


(fl 




(fl 
a) 


,_ ) 


1— 


3 

0) 


TO 


^ 


T3 


"□ 


i- 




O 


^ 


5 


o 
>- 


u 

£ 


■5 


■l-l 


0) 


r 


^ 


o 






o 


(A 


^ 


§ 


o 

r 


o 

T3 


-I 


C 


< 


g 


o 


CO 


LL 


O) 


CU 


O) 


J_ 


rH 


H 








O 


O) 


C 


1 


o 


CM 


in 


a) 






b 


"5 


fl) 


TO 


LL 

r 


LL 






g 




T3 




ri) 




() 




1 




T5 




fl 




!— 




n 




CI) 




tr 




cu 




o 




!~ 




3 




O 




w 





26 



N/ST NCSTAR 1-7, WTC Investigation 



Design of the World Trade Center Egress System 



A similar condition existed on the 107' floor of WTC 2, commonly referred to as the Observation Deck. 
A tenant alteration application submitted by Ogden Entertainment (the tenant) to PANYNJ in late 1 995 
and early 1 996 utilized the areas of refuge provisions referred to previously with respect to the Windows 
on the World space. ' Taking advantage of a NYC Building Code provision which permits a lower basis 
for occupant load, the PANYNJ permitted a maximum occupant load of 1,170 persons on the floor 
(Indoor Observation Deck and Outdoor Observation Deck, combined), which was enforced by the lessee 
with periodic oversight by PANYNJ. 12 



2.2.2 



Stairwells 



WTC 1 and WTC 2 each had three 

primary stairwells designed for 

emergency egress, designated as A, B, 

and C. There were additional 

stairwells located in the basement 

levels (B 1 - B5), convenience stairs 

for tenants leasing multiple floors, 

and mechanical room stairs. These 

secondary stairs are not considered 

part of the emergency egress system 

and are not described here. Stairwells 

A and C were 1.1m (44 in.) wide and 

extended from floor 2 (plaza or 

Mezzanine Level) to floor 110 (lower 

mechanical space). The stairwell 

landings by the exit door were 92 in 

(2.3 m) wide by 78 in (2.0 m) deep. 

Figure 2-10 shows a 44 in. (1.1 m) 

stairwell in WTC 1 taken on September 

11, 2001, by John Labriola during his evacuation. Note the photo luminescent paint on the stair edge and 

landing. Stairwell B was 56 in. (1.4 m) wide and ran from the subgrade 6 levels below ground to floor 

107 including the Concourse (main lobby); there was no exit from Stairwell B onto the 2nd floor (plaza / 

Mezzanine Level). The stairwell landings by the exit door for Stairwell B were 1 16 in (2.9 m) wide by 78 

in (2.0 m) deep. 

The 1968 NYC Building Code has requirements for the number and capacity of stairs and for the assumed 
occupant load that are similar to requirements in the other contemporaneous codes (see NIST 
NCSTAR 1-1, Appendix A). Codes of that time required that multiple stairs be located "as remote from 
each other as practicable." NYC permitted scissor stairs, 1 and the code required the exit doors to be at 
least 4.6 m (15 ft) apart. Local Law 16 (1984) first imposed a remoteness requirement of 30 ft or one- 




Figure 2-10. 44 in. stairwell in WTC 1 taken on 
September 11, 2001. 



Ogden Entertainment. 1996. Port Authority work number W96-2103-01. WTCI-180-P. 
12 PANYNJ response to formal NIST question. "Re: Question for PA." March 25, 2005. S. Bohl to S. Sunder. 

Scissor stairs are two separate stairwells with two separate stairwell access doors, which share a common shaft space, often 
winding around each other. This results in an efficient use of space, but places the stairwells in direct contact (in other words, 
there is not a barrier separating the stairwells), thus allowing smoke or other threats to affect two stairwells simultaneously. 



NIST NCSTAR 1-7, WTC Investigation 



27 



Chapter 2 



third the maximum travel distance of the floor (whichever is greater). This requirement was not 
retroactive, so it did not apply to WTC 1 and WTC 2. However, this requirement did apply to WTC 7. 

The 1968 NYC Building Code also states that, " ...vertical exits should extend in a continuous enclosure 
to discharge directly to an exterior space or at a yard, court, exit passageway or street floor lobby ..." 
(C26-602.4). The 1965 BOCA Basic Building Code and 1966 NFPA 101 contained similar language, but 
not the 1964 New York State Building Construction Code or the 1966 Municipal Code of Chicago. 
Current model code language (2003 IBC, section 1003.6) defines continuous as: not "... interrupted by 
any building element other than a means of egress component." 

The exit discharge language was the subject of discussion in that the stairs in WTC 1 and WTC 2 
discharged onto the Mezzanine Level, which was not at street level but rather at the Plaza level. The Port 
Authority took the position that the concourse was like an underground street, and the arrangement met 
the intent of the Code, as demonstrated by a February 18, 1975 letter from Joseph Solomon (Emory Roth 
and Sons) to Malcolm Levy (PANYNJ), which covered six points. "We [Emory Roth and Sons] were 
instructed by the Port Authority to deviate from the code [1968 NYC Building Code]." The fourth point 
listed the "treatment of concourse level as 'Underground Street' noted by letter to the Port Authority on 
April 6, 1971, January 11, 1972, and May 7, 1973" (Solomon 1975). 

Transfer Hallways 

The WTC 1 and WTC 2 stairwells were occasionally routed horizontally around equipment on 
mechanical floors, through what were called transfer hallways, as shown in Figure 2-11. Table 2-2 
shows the overall layout of the stairwells in WTC 1 and WTC 2, including the basic core perimeter. 14 
Stairwell B required a horizontal transfer at floor 76. For all other floors, stairwell B maintained vertical 
alignment through the building. Stairwells A and C required horizontal transfers (some longer than 
others) at floors 42, 48, 66, 68, 76, and 82. Horizontal transfer distances ranged from several feet 
(floors 66 and 68) to over 100ft (33 m), including smoke doors (which were closed but not locked) and 
multiple right angles turns in the transfer on floors 42, 48, 76, and 82 for Stairwells A and C. Note that 
the mechanical floors were located on floors 41-42, 75-76, and 108-109. One problem with the horizontal 
transfers was that they extended the total evacuation time, when compared to a similar design without 
horizontal transfers. The World Trade Center Review Committee, formed by the New York City 
Building and Fire Commissioners in response to the 1993 WTC Bombing, found that "the occupants of 
the towers encountered changes in the path of egress that were unfamiliar, [contributing] to the general 
confusion during the evacuation process (New York City 1995)." Figure 2-12 shows a photograph of a 
horizontal transfer hallway in WTC 1 or WTC 2 taken after the 1993 bombing, including 
photoluminescent markings. 



Core is defined in this report as the boundary of non-leasable common space, including egress hallways, stairwells, elevator 
shafts and lobbies, HVAC, plumbing and other mechanical spaces. This definition of core may differ from a structural 
definition of core, defined by the location of core columns, which did not change location in WTC 1 or WTC 2 on different 
floors. 

28 NIST NCSTAR 1-7, WTC Investigation 



Design of the World Trade Center Egress System 




WTC 2 



Figure 2-11. Stairwells in the WTC towers. 



WTC1 



NISTNCSTAR 1-7, WTC Investigation 



29 



Chapter 2 




Each stairwell had signage on both 
sides of the stairwell access doors 
indicating the letter designation of the 
particular stairwell. A sign on the 
inside of the stairwell indicated the 
floor number, the stairwell 
designation, and whether the floor was 
a "re-entry" or "non-re-entry" floor. 
Figure 2-13 shows a photograph of 
this signage taken after the 1 993 
bombing. A non-re-entry floor was a 
landing in the stairwell where the door 
to the floor was locked from the 
stairwell side. If the particular floor 
was not a re-entry floor, the sign 
indicated the location of the nearest re- 



Figure 2-12. Horizontal transfer floors in the 
WTC towers. 

entry location, every fourth floor (in the case of 
Figure 2-13, floors 74 and 78). The stairwell doors 
were required to be always open every fourth floor by 
the NYC Building Code. Door locks leading to 
mechanical spaces and the roof were controlled 
electronically at the Security Command Center (SCC) 
on floor 22. The NYC Building Code also required 
that, in the event of a power outage, the re-entry 
locking mechanism would default to the open position. 

Compartmentation 

The design of WTC 1 and WTC 2 featured large, open 
office spaces devoid of columns due to the innovative 
structural design. Tenants could (and often did) utilize 
open plan office layouts that permitted impressive 
views of the Manhattan skyline out the perimeter 
windows. 



The NYC Building Code and PANYNJ practice 

required partitions to separate tenant spaces from one 

another and from common spaces such as the corridors 

that served the elevators, stairs, and other common 

spaces in the building core. Fire rated partitions are 

intended to limit fire spread on a floor and to prevent 

spread of fire in one tenant space to that of another. 

Partitions separating tenant space from exit access 

corridors were permitted to be 1 h, although PANYNJ 

specified them to be 2 h (Kyle 1966). This allowed dead ends to extend to 100 ft (rather than 50 ft with 

1 h partitions), which permitted more flexibility in tenant layouts. Partitions separating tenant spaces (so 




Figure 2-13. Stairwell door signage as 
seen from inside stairwell A. 



30 



NIST NCSTAR 1-7, WTC Investigation 



Design of the World Trade Center Egress System 



called demising walls) were required to be 1 h. Enclosures for vertical shafts, including stairways and 
transfer corridors, elevator hoistways, and mechanical or utility shafts were required to be of 2 h fire rated 
construction. Protection of vertical shafts was intended to limit the spread of fire and smoke from floor to 
floor. 

Another influence on compartmentation of the buildings was the adoption of Local Law 5 
(New York 1973) (LL 5) amending the NYC Building Code. While it did not (legally) apply to the 
WTC buildings, PANYNJ policy was to follow the requirements voluntarily. LL 5 required 
compartmentation of unsprinklered spaces in existing office buildings over 100 ft in height "having air- 
conditioning and/or mechanical ventilation systems that serve more than the floor on which the equipment 
is located" to be subdivided by 1 h fire separations into spaces or compartments not to exceed 7,500 ft 2 . 
Floor areas could be increased up to 15,000 ft 2 if protected by 2 h fire resistive construction and smoke 
detectors. Regardless of the floor area, compartmentation is not required when complete sprinkler 
protection is provided (LL 5, Section 6). 

Shortly after the adoption of LL 5, PANYNJ began to add the required compartmentation as a part of new 
tenant layouts as evidenced by several subsequent tenant alteration contracts from this time. Following 
the 1975 fire a fire safety consultant report recommended to PANYNJ that the buildings be retrofitted 
with sprinklers to address possible smoke problems, which would also obviate the need for 
compartmentation and permit the unobstructed views for which the buildings were known. The decision 
left the interior WTC floor arrangements with only partitions separating tenant spaces from one other and 
from exit access corridors or common spaces in the core, and with shaft enclosures. 5 

Construction of Partitions and Shaft Enclosures 

Vertical shafts surrounding stairs, mechanical shafts (carrying supply and return air), elevator hoistways, 
and utility shafts were all contained within the building core of the WTC towers, and were enclosed by 
gypsum planking similar to fire separations commonly used today in single-family attached housing. 
These gypsum planks were 2 in. thick and 2 ft wide, reportedly with metal tongue and groove channels 
attached to the long sides. These were likely two 1 in. panels held together by the metal channels. Their 
length in WTC 1 and WTC 2 is unknown, but similar panels today are available in 8 to 14 foot lengths. 
The planks were placed into metal H-channels at the top and bottom and secured by drywall screws. 

The 1978 edition of the Gypsum Association (GA) Fire Resistance Design Manual lists several similar 
shaft wall constructions utilizing 2 in. gypsum planks consisting of two 1 in. gypsum core board panels 
with "metal channels on long edges." The GA Manual lists shaft walls of a single 2 in. metal edged plank 
(WP7015) having a 1 h fire rating, a single 2 in. metal edged plank with one layer of Type X gypsum 
board on the unexposed side (WP71 12) having a 2 h fire rating, and a single 2 in. metal edged plank with 
two layers of Type X gypsum board on the unexposed side (WP 7575) having a 3 h fire rating. 

Partitions separating tenant spaces on the same floor were constructed of two layers of 5/8 in. Type X 
gypsum board on steel studs and ran slab to slab. This construction is commonly recognized as a 2 h fire 
separation. Above the ceiling, penetrations for ducts or to allow for return air flow were fitted with rated 
fire dampers to preserve the fire rating. 



15 PONYA 1976 - Complete report reproduced in NIST NCSTAR 1-1H. 

NIST NCSTAR 1-7, WTC Investigation 3 1 



Chapter 2 



Interior partitions not separating spaces occupied by different tenants were constructed of single or double 
layers of 5/8 in. Type X gypsum board on steel studs, and ran from the slab to the suspended ceiling but 
not above. Double layers of gypsum board were used when the tenant desired additional sound 
attenuation. These partitions were not required to be fire rated and fire rated doors were not used. 
However, a single layer of 5/8 in. Type X gypsum board on steel studs (16 in. on center) is generally 
considered to have a 1 h fire rating and two layers of 5/8 in. Type X gypsum on steel studs (16 in. on 
center) is considered to have a 2 h fire rating. For a ceiling-high partition to be considered as having a 
fire rating, the ceiling itself would have to be rated as well. The ceiling system used throughout these 
buildings was not fire rated. 

2.2.3 Elevators 

Getting thousands of people from the ground level to the offices, observation levels, and restaurants, some 
as high as a quarter- mile, was no small task. Thus, elevators were the primary mode of movement 
between floors of the World Trade Center. The World Trade Center complex contained more than 
240 elevators, with 99 elevators serving the above-ground levels in each of the two main towers and an 
additional 7 elevators serving primarily the sub-grade basement levels. In the towers, the elevators were 
arranged to serve the buildings in three sections divided by skylobbies, which served to distribute 
passengers among express and local elevators. Figure 2-14 shows an elevator riser diagram for WTC 1 
and WTC 2 for passenger elevators. 

• People traveling to floors 9 through 40 entered a bank of 24 local elevators at the Concourse 
Level. These were divided into four groups, with each stopping at a different set of eight or nine 
floors (9 through 16, 17 through 24, 25 through 31, and 32 through 40). 



• 



• 



Those going to floors 44 through 74 took one of eight express elevators to the 44th floor skylobby 
before transferring to one of 24 local elevators. These 24 were stacked on top of the lower bank 
of 24, providing additional transport without increasing the occupied floor space. 

Those going to floors 78 through 107 took one of 10 express elevators from the Concourse Level 
to the 78th floor before transferring to one of 24 local elevators. These were also stacked on the 
lower banks of 24. 

Dedicated express elevators served the restaurant, bars, and meeting rooms on floors 106 and 107 
of WTC 1, as well as the observation deck in WTC 2. 

An occupant traveling to the 91st floor, for example, would have taken an express elevator from the lobby 
to the 78th floor and then would have had to transfer to another elevator to arrive at the 91st floor. The 
elevator trip would have taken several minutes travel time, depending upon the wait at the elevators. 
While providing an acceptable rate of people movement, this three-tier system also used less of the 
building footprint than the usual systems in which all elevators run from the entrance to the top of the 
building. Further, leasable floor space was reclaimed near the top of a given zone. At the top of each 
elevator bank, the machinery to lift the cabs occupied the next higher floor. From the next higher floor up 
to the bottom of the next elevator bank, there was no need for an elevator shaft. The concrete floor was 
extended into this space, providing additional rentable floor area for offices, conference rooms, storage, 
etc. Figure 2-14, for example, shows that the space taken by Elevator Bank A (Elevators 24 - 29) in 



32 NIST NCSTAR 1-7, WTC Investigation 



Design of the World Trade Center Egress System 



order to serve floor 9 to floor 16, was reclaimed for tenant use on floors 19 to 42. This resulted in a 
reclamation of approximately 750 ft 2 per floor. A calculation for reclaiming unused floor space above 
elevator banks A, B, and C for all three zones, reveals that roughly 100,000 ft 2 of potentially leasable 
office space could be recovered. Assuming $55 per ft 2 per year as a rental rate for a downtown 
Manhattan office building over 600,000 ft 2 (BOMA 2001), the reclamation could theoretically yield 
nearly $6 million per year of rental income. At the time WTC was built, the concept of skylobbies, 
served by express elevators and serving only one zone of the building, was innovative. Other tall 
buildings now use this concept. 



Mechanical Equipment -p. 
Room 



Skylobby -•_ 

Mechanical Equipment -• 

Room = 



Skylobby -J| 

Mechanical Equipment 
Room 



Express Elevators - = 



Mechanical Equipment 
Room 

Plaza Level 



: = 



Local Elevators 



Express Elevators 



Local Elevators 



Figure 2-14. Elevator riser diagram for WTC 1 and WTC 2. 



NISTNCSTAR 1-7, WTC Investigation 



33 



Chapter 2 



In addition to the passenger elevators, there were seven freight elevators in each tower; most served a 
particular zone, while Car 50 served every floor. 

Car #5: B 1-5, 7, 9-40, 44 

Car #6: Bl-5, 44, 75, 77-107 (Dual-use express, see below) 

Car#17:Bl-l, 41, 43-78 

Car#48:Bl-7, 9-40 

Car #49: Bl-5, 41-74 

Car#50:B6-108 

Car #99: 107-1 10 16 

There were two express elevators (#6 and #7) to Windows on the World (and related conference rooms 
and banquet facilities) in WTC 1 and two to the observation deck in WTC 2. There were five local 
elevators in each building: three that brought people from the subterranean levels to the lobby, one that 
ran between floors 106 and 110, and one that ran between floors 43 and 44, serving the cafeteria from the 
skylobby. All elevators had been upgraded to incorporate firefighter emergency operation requirements. 

Local Law 5 (New York 1973) requires that elevators be provided with an emergency recall system. This 
requirement was incorporated subsequently into the American Society of Mechanical Engineers (ASME) 
A17.1, Safety Code for Elevators and Escalators, which governs elevator design and operation in all 
present U.S. building codes. The ASME Code required that: 

• All passenger elevators be marked with signs stating that they cannot be used during a fire; 

• Fire detectors installed in every elevator lobby and machine room be arranged to initiate a recall 
of the elevators to the ground floor where the doors open and the elevator is taken out of service; 
and 

• Fire service personnel can use a special key to operate any individual car in a manual mode as 
long as they feel it is safe to do so. 

• At least one elevator serving every floor be connected to emergency power. 

Currently, there are no national model codes that permit elevators to be used as a means of occupant 
egress in emergencies, and national standard ASME A17.1 (ASME 2000) requires signs at all elevators 
warning that they should not be used in fires. There are some recent exceptions to this requirement, but 
these are limited to special cases. For example, NFPA 5000 permits protected elevators as a secondary 
means of egress for air traffic control towers, and the City of Las Vegas accepted elevators as a primary 
means of occupant egress from Stratosphere Tower based on a performance-based design (Bukowski 
2003). 



16 PANYNJ WTC Fire Safety Plan (Towers One and Two), 1995, WTCI-13-NYC. 

34 NIST NCSTAR 1-7, WTC Investigation 



Design of the World Trade Center Egress System 



U.S. building codes (including NYC Building Code) require accessible elevators as part of a means of 
egress that may be used by the fire service to evacuate people with disabilities. These elevators must 
comply with the emergency operation requirements of ASME A17.1 (Phase II emergency operation by 
the fire service), be provided with emergency power, be accessible from an area of refuge or a horizontal 
exit (unless the building is fully sprinklered), and operate in a smoke protected hoistway. Phase II 
operation involves the use of an elevator by a firefighter for fire service access or for rescue of people 
with disabilities performed under manual control (with the use of a special key). 

In the event of a fire in WTC 1 or WTC 2, or other emergency requiring evacuation where the stairwells 
are unusable or cut off by fire and/or smoke, consideration of using elevators for occupant egress may be 
given in accordance with the following PANYNJ guidelines: 

• Elevators may not be used if they also service the fire floor, except under specific instructions 
from the fire safety director or Fire Department; 

• If the elevators do not service the fire floor and their shafts have no opening to the fire floor, they 
may be used at the direction of the fire safety director or fire department; 

• Elevators under the direction of the fire department or trained building personnel may be used. 7 

Every elevator lobby contained a sign reading, "IN CASE OF FIRE USE STAIRS UNLESS 
OTHERWISE INSTRUCTED." The sign also included a diagram indicating the location of the sign and 
the location and letter designation of each stairwell serving the particular floor. 18 

2.2.4 Emergency Communication System 

WTC emergency procedures specified that all building-wide announcements were to be broadcast from 
the fire command station of each WTC tower, in coordination with the fire safety director or life safety 
and security supervisor. The deputy fire safety director was likely to make all announcements. 
Appendix J of the World Trade Center Emergency Guidelines 19 provided prepared text for a variety of 
emergency scenarios, including power failures, fires, and service interruptions. Prior to all emergency 
announcements, the following pre-announcement was made: 

"Your attention please, your attention please. An important public 
address announcement will be made in the main corridor of your floor in 
a few moments." 

Evacuation for any reason, including fire or smoke, would have generated the following announcement, 
enabling a phased evacuation: 

"Your attention please. We are experiencing a smoke condition in the 
vicinity of your floor. Building personnel have been dispatched to the 
scene and the situation is being addressed. However, for precautionary 
reasons, we are conducting an orderly evacuation of floors . 



17 World Trade Center Emergency Procedures Manual 2001 - Confidential. Port Authority of New York and New Jersey. 

18 The Port Authority of NY & NJ World Trade Center Fire Safety Plan (Towers One and Two). 1995. WTCI-13-NYC. 
World Trade Center Emergency Procedures Manual 2001 - Confidential. Port Authority of New York and New Jersey. 

NIST NCSTAR 1-7, WTC Investigation 35 



Chapter 2 

Please wait until we announce your floor number over the public address 
system. Then follow the instructions of your fire safety team. We will 
continue to keep you advised. We apologize for the inconvenience and 
we thank you for your cooperation." 

The standard evacuation announcement for a particular floor was: 

"Your attention please. It is now time for your floor to be evacuated. In 
accordance with the directions from your fire safety team, please take the 
exit stairs nearest to your location. We remind you that communications, 
emergency lighting and other essential services are in service. We will 
continue to keep you advised. We apologize for the inconvenience and 
we thank you for your cooperation." 

According to the Guidelines, however, the information and instructions broadcast to the building 
occupants could be modified to suit the nature of the emergency, at the discretion of the fire safety 
director. 20 NIST NCSTAR 1-4 addresses the fire alarm systems in WTC 1 and WTC 2. 

Fire Command Station 

The fire command station, located in the lobbies of both WTC 1 and WTC 2, provided a command post 
for building personnel to orchestrate the response. The NYC Building Code requires that the computer 
screen in the fire command station monitor and display information regarding: 

Manual fire alarms 

Smoke detection 

Sprinkler water flow 

Elevator lobby smoke detectors 

Fire signal activation 

Central office notification 

Fan system status 

Fail safe locked door status 



20 



World Trade Center Emergency Procedures Manual 2001 - Confidential. Port Authority of New York and New Jersey. 
36 NIST NCSTAR 1 - 7, WTC Investigation 



Design of the World Trade Center Egress System 



Fire system trouble 

Fire signal trouble 

Tamper switch alarm 

Power source 

Test/normal mode 

Other information as desired, 
including the status of 
elevators. 



The primary value of the fire 

command station was its role as a 

convening point for key building 

personnel responding to a building 

incident. The roles of many of the 

key personnel are described in the 

following section. Figure 2-15 shows 

the fire command station in the lobby of WTC 1 on September 11, 2001, seen from the east end of the 

Mezzanine Level. The fire command station appears in the back right corner of the picture. 




Figure 2-15. Fire command station in lobby of WTC on 

September 11, 2001, as seen from mezzanine 

escalator, looking west. 



2.3 



THE HUMAN ELEMENT 



PANYNJ produced and regularly updated an emergency procedures manual for building personnel to 
follow in the event of a building incident, 21 at least until Silverstein Properties formally had become 
leaseholder several months prior to September 11, 2001. While Silverstein Properties was formally 
managing WTC 1 and WTC 2, PANYNJ staff continued to be significantly involved in property 
management during the transition. The latest update to the manual was completed earlier in 2001. Note 
that PANYNJ was not responsible for responding to fires or alarms in WTC 3 (Marriott Hotel), WTC 6 
(US Customs House), or WTC 7. 

The fourteen chapters in the 2001 manual addressed such possibilities as bomb threats, fires, floods, gas 
leaks, elevator emergencies, power failures, medical emergencies, chemical and fuel releases, structural 
integrity, and political demonstrations, among other potential problems. Aircraft impact was not 
specifically addressed. Individual responsibilities for key personnel were enumerated, including 
interactions with non-PANYNJ personnel (including, as appropriate, FDNY, NYPD, and others). The 
following subsections of Chapter 2.3, are derived primarily from The 2001 WTC Emergency Procedures 
Manual, Chapter 7, entitled "Fire Emergencies." 



World Trade Center Emergency Procedures Manual 2001 - Confidential. Port Authority of New York and New Jersey. 



NISTNCSTAR 1-7, WTC Investigation 



37 



Chapter 2 



2.3.1 Responsibilities of the Fire Safety Director 

The fire safety director was a position required by Local Law 5 (New York 1973). Local Law 5 required 
all buildings (new or existing) occupied by more than 100 persons above or below the street level, or 
more than 500 people in the entire building to have a fire safety director, deputy fire safety director, and 
building evacuation supervisor. Each such building is required to have one employee designated as fire 
safety director and one or more employees designated as deputy fire safety director, who possess 
certificates of fitness from the commissioner qualifying the individual to conduct fire drills, evacuations, 
and related training. A certified individual is required to be on duty during normal working hours. 
Consistent with Local Law 5, the primary responsibility of a fire safety director at the WTC (according to 
the formal emergency procedures manual) was overall emergency management for a building incident 
(PANYNJ 2001b). The fire safety director reports to the Fire Command Station, or scene, and assumes 
the following duties: 

Verify that FDNY has been notified and coordinate activities of FDNY and other emergency 
response personnel; 

Confer with floor wardens of affected floor(s) to determine conditions on the floor and identify 
areas to be evacuated, route of evacuation, stairwells available, and potential refuge floors; 

Initiate evacuation procedures; 

Direct public address announcement(s), as necessary; 

Deploy security officers to restrict access to affected and secure areas; 

Dispatch "key runs"; 22 

Ensure appropriate notifications are initiated; 

Maintain a chronological record of the event; 

Direct the Operations Control Center (OCC) to arrange for emergency elevator service; 

Investigate cause of fire (in coordination with the FDNY Bureau of Fire Investigation, prepares 
appropriate reports). 

2.3.2 Responsibilities of the Deputy Fire Safety Director 

The role of the deputy fire safety director assigned to the Fire Command Station in the lobby was to 
execute and direct the fire safety plan, including: 

• Notify fire safety director, operations control center, and police desk (WTC 5) of incident; 

• Maintain communications with floor wardens and other members of the fire safety team; 



A "key run" is a security officer charged with distributing and retrieving master key rings and Fire Department portable radios. 
These keys provided access to secure areas for emergency responders or building personnel during an incident. 

38 NIST NCSTAR 1 - 7, WTC Investigation 



Design of the World Trade Center Egress System 



• Assist with crowd control and evacuation, as necessary; 

• Request that the elevator starter or OCC contact the elevator(s) to respond to the Lobby and await 
Fire Department personnel; 

• Complete necessary fire alarm notification forms. 

2.3.3 Responsibilities of the Operations Control Center Supervisor 

Upon notification of a fire event, the supervisor on duty at the Operations Control Center was to first 
ensure that the fire command station and the fire safety director were notified. Next, the supervisor was 
to issue a general broadcast of information over all WTC radio channels, monitor all channels and ensure 
that radio silence is observed unless directly related to the ongoing incident, arrange for elevator service, 
update units with relevant information as necessary, and notify managers of Windows on the World 
(WTC 1) and Top of the World (WTC 2) of incident in order to "reduce anxiety to tenants, visitors, 
guests, etc. when numerous emergency vehicles respond." The Operations Control Center was located in 
the Bl Level of WTC 1 and was a backup Fire Command Center. 

2.3.4 Responsibilities of the Operations and Maintenance Management 

Building operators and maintenance personnel were mobilized in order to provide emergency response 
assistance should the need arise. The duty supervisor established contact with the fire safety director, fire 
safety coordinator, or life safety and security supervisor and responded to the fire command station to 
assist as required. The operations group supervisor, who may have required self contained breathing 
apparatus, was assigned to respond to one floor below the scene of the incident, established 
communication with the fire command station using the floor warden telephone, assisted with the 
evacuation, and kept in contact with the fire command station. 

The supervisor of the mechanical contractors was to dispatch staff to the fire pumps in order to "stand by" 
for further instructions, dispatched staff to operate the smoke purge system as requested by the fire safety 
director or Fire Department, and dispatched staff to secure sprinkler water shutoff valves. 

The supervisor of electrical contractors was assigned to dispatch one contract electrician to one floor 
below the affected floor in order to assist should the incident involve electrical closets or fixtures, two 
electricians to the nearest sub-station below the affected floor, and a supervisor to the fire command 
station. Further, the electrical supervisor was to ensure that staff was standing-by in order to secure 
electrical power, if necessary, and that portable electrical power was available, as needed, and played a 
significant role in post-incident restoration of smoke detectors and/or alarm panels. In the event of a 
major disaster, all staff electricians were to report to the electrical shop/office. 

The elevator maintenance contract supervisor was to report to the fire command station in order to assist, 
as needed, as well as dispatch elevator mechanics to their appropriate posts to assist, as needed. 
Figure 2-16 shows a WTC official (denoted by the vest identifying WTC Officials) attempting to 
communicate with elevator occupants in WTC 1 on September 11, 2001, from the fire command station 
in the lobby. 



NIST NCSTAR 1-7, WTC Investigation 39 



Chapter 2 




2.3.5 



Figure 2-16. Elevator communication panel in the fire 

command station of WTC 1, as operated on 

September 11, 2001. 

Floor Warden System 



The WTC Emergency Procedures (PANYNJ 2001b) requires each floor of a high-rise building to 
designate a floor warden to coordinate the evacuation of the floor, consistent with NYC Building Code. 
Assisting the floor warden were deputy floor wardens and searchers, which constitute a tenant fire safely 
team. On multi-tenant floors, each tenant identified a floor warden for their space. Once the order to 
evacuate a floor was given, those with building authority had specific responsibilities to insure an orderly 
evacuation: 

• In the event of an emergency, the floor warden was responsible for ensuring that an alarm was 
transmitted by either telephoning the police desk or activating a manual pull station. The floor 
warden reported the incident in detail to the Fire Command Station, and relayed instructions to 
building occupants. 

• The floor warden was responsible for notifying occupants of the floor that there was a fire and 
ensure that the occupants executed the fire safely plan (PANYNJ 1995). In an emergency, 
searchers would round up employees, and the deputy fire warden would move them into the 
corridors and make sure all occupants were accounted for. In the event occupants were reluctant 
to evacuate, searchers were not required to force evacuation. 

• In coordination with the Fire Safety Director, floor wardens selected the safest stairwell to use on 
the basis of the location of the fire, including checking the environment in the stair, and notifying 
the fire command station which stairwell was utilized. 



2.3.6 



Occupant and Tenant Training 



23 



WTC policy was to conduct fire drills every 6 months, consistent with NYC Local Law 5, or shortly 
after move-in for all new tenants in WTC 1 and WTC 2. Written procedures specified a three day 



Local Law 5-73. §C19-161.2.a.4. 



40 



NIST NCSTAR 1-7, WTC Investigation 



Design of the World Trade Center Egress System 



advance notice prior to the drill for tenants, through the floor warden and deputy floor warden. The floor 
warden then notified all occupants of the floor. 

Immediately prior to the fire drill, the public address system would be used to announce that the drill was 
about to occur. Occupant attendance at drills was mandatory, with a small "skeleton staff permitted for 
business continuity. An occupant who missed a fire drill as "skeleton staff was required to attend the 
next fire drill. The occupants were required to assemble outside a designated stairwell. 

During the fire drill training, the fire alarm was sounded. The floor warden, deputy floor warden, and 
searchers ensured that occupants gathered in the central hallway, near a stairwell. The fire safety team 
then instructed the occupants not to attempt to fight fires, not to use the elevators, to obey all instructions 
from the deputy fire safety director, and what phone number to call if there was a problem. The location 
of the nearest stairwell was identified and the procedures for phased-evacuation (move three floors below 
the fire floor, as instructed by the floor warden and/or deputy fire safety director) (PANYNJ 1996). 

The standard instruction to the occupants was to evacuate downward (to three floors below the incident 
floor). The training did not explicitly instruct occupants not to evacuate upward or attempt to access the 
roof. Stairwells A and C went to the 1 10th floors, but only to serve as egress points to descend from the 
1 10th floor or the roof. The 1 10th floor was not a re-entry floor, and thus, occupants without an 
authorized badge or a key would have been unable to reach the door that led to the roof. Had the 1 1 0th 
floor been accessible, actually reaching the roof would have been prevented by two additional doors, in 
accordance with Federal Communication Commission regulations. 24 The first door to access the stairwell 
to the roof was protected by an access card reader. Upon opening the first door, the individual would 
enter a vestibule where, upon showing ID to a closed-circuit television monitored at the Operations 
Control Center (OCC), the door would be electronically unlocked from the OCC. Access to the roof was, 
thus, limited to a small number of people certified to enter through a radio frequency hazard awareness 
class. 24 

Floor wardens, deputy floor wardens, and searchers were required as part of their training, to watch a 
video, prepared by PANYNJ. The video entitled "WTC Fire Safety" and provided to NIST by PANYNJ, 
reviewed the emergency procedures, building fire safety systems, and the responsibilities of the members 
of the fire safety team (PANYNJ 1996). 

2.4 CHANGES TO THE EGRESS SYSTEM AFTER THE 1993 BOMBING 

The February 26, 1993, World Trade Center bombing precipitated a $250 million 25 repair and life safety 
upgrade to the complex, including (PANYNJ undated): 

• Radio repeaters on the roof of WTC 5 for Fire Department communications. 

• Circulation improvements. 



The roof housed critical communications equipment, including broadcast facilities for major television stations, paging 
transmitters, FDNY transmitters, numerous mobile transmitters. [Information derived from Port Authority response to formal 
NIST question, April 2004] 
25 Public Comments to Draft Version of NCSTAR 1-7. PANYNJ. 2005. 

NIST NCSTAR 1-7, WTC Investigation 41 



Chapter 2 



New North (to Vesey St.) and South (to Liberty St.) corridors for faster evacuation from the 
Concourse (mall). 

Two escalators from the Concourse (mall): one to the plaza at WTC 5 and one up to WTC 4 and 
onto Church St. 

Photoluminescent paint on handrails, stair treads, and stair centerline. 

Multiple power sources for stairwell lighting: 2 normal feeds, back-up generator, and a back-up 
from the PATH system; battery backup for every other stairwell fixture (up to 90 minutes). 

LED exit signs for extra brightness and visibility through smoke conditions. 

Fluorescent signs inside stairwells at all stair reentry doors along with raised porcelain type 
Braille. 

Fire Command Stations in main lobbies. 

Two sealed beam (with battery back-up) elevator lights and bells, in addition to normal lighting. 

Upgraded elevator intercom system, monitored at Fire Command Station. 

New decentralized Fire Alarm System (Style 7), with three separate data risers to transponders 
located every three floors; redundant control panels and electronics; multiple control station 
announcement capability. 

Fire alarm system powered by normal emergency power, battery back-up, and tertiary power to 
equipment. 

New modernized Operations Control Center with the capability to monitor all HVAC systems 
and elevators. 

Elevators modernized to current code, including replacing relay system with microprocessor 
based system (only 50 percent complete on September 11, 2001). 

Sprinkler installation accelerated to completion, including Concourse. 

Fire wardens equipped with flashlights, whistles, hats and special training. 

Fire drills conducted in conjunction with the Fire Department. 

In addition, PANYNJ purchased evacuation chairs, which were provided free to building tenants upon 
request, for use by mobility-impaired occupants during emergency evacuation. Further discussion of the 
changes to WTC 1 and WTC 2 subsequent to the 1993 bombing can be found in NIST NCSTAR 1-1. 



42 NIST NCSTAR 1-7, WTC Investigation 



Design of the World Trade Center Egress System 



2.5 BUILDING CODE ANALYSIS 

NIST NCSTAR 1-1 addresses the building codes relevant to WTC, including provisions for egress system 
design. For most buildings constructed in the United States, building codes adopted by local jurisdictions 
establish minimum requirements for design and construction. However, because the PANYNJ is an 
interstate agency, which was established in 1921 under a clause in the U.S. Constitution, its construction 
projects are not required to comply with any local or national model building code. For the design of the 
WTC towers, which began in 1962, the Port Authority in May 1963 instructed the architect and engineers 
to prepare their designs of WTC 1 and WTC 2 to comply with the NYC Building Code. 26 While not 
specifically stated in the 1963 letter to the architect, the 1938 edition of the Code was in effect at that 
time. In areas where the Code was not explicit or where technological advances made portions of the 
1938 Code obsolete, the Port Authority also directed the architect and engineers to propose designs 
"based on acceptable engineering practice." When such situations occurred, the Port Authority required 
the architect and engineers to inform the Planning Division of the WTC. The Port Authority established a 
special WTC office that reviewed and approved plans and specifications, issued variances, and conducted 
inspections during construction instead of the city agencies that would normally perform these duties. 

In September 1965, the Port Authority instructed the architect and engineers to revise their designs for 
WTC 1 and WTC 2 to comply with the second and third drafts of the NYC Building Code that was under 
development and to undertake any design modifications necessary to comply with the new code 
provisions. 27 Prior to issuance of this instruction, the Port Authority recognized that the draft version of 
the new New York City Building Code had incorporated advanced techniques and that the Port Authority 
favored the use of advanced techniques in the design of the WTC towers. 28 By adopting the draft versions 
of the new NYC Building Code, WTC 1 and WTC 2 were classified as Type 1-B Construction instead of 
Type 1-A Construction (see Sect. 9.1.3 for definition and fire protection requirements of Construction 
Type), and several architectural features related to egress were modified in the final design (see Sect. 10.1 
of NIST NCSTAR 1-1). This relaxation of code requirements allowed the Port Authority to gain 
economic advantage. 29 The new NYC Building Code (NYCBC 1968) was enacted by the City Council 
on October 22, 1968, approved by the Mayor on November 6, 1968, and became effective on 
December 6, 1968. 

2.5.1 Egress in the Building Codes 

The ability to evacuate thousands of occupants from buildings as massive as WTC 1 and WTC 2, was a 
function of three primary variables: how many stairs, how wide the stairs were, and where the stairs were 
located. Each of those three factors, in the context of building code requirements, are evaluated below. 



26 Letter dated May 15, 1963 from Malcolm P. Levy (Chief, Planning Division, World Trade Department, PANYNJ) to Minoru 
Yamasaki (architect, Minoru Yamasaki & Associates) (See Appendix A of NCTAR 1-1). 

27 Letter dated September 29, 1965 from Malcolm P. Levy (Chief, Planning Division, World Trade Department, PANYNJ) to 
Minoru Yamasaki (architect, Minoru Yamasaki & Associates) (See Appendix A of NCSTAR 1-1). 

28 Memorandum dated June 22, 1965 from John M. Kyle (Chief Engineer, PANYNJ) to Malcolm P. Levy (Chief, Planning 



29 



Division, World Trade Department, PANYNJ) (See Appendix A of NCSTAR 1-1) 

Memorandum dated January 15, 1987 from Lester S. Feld (Chief Structural Engineer, World Trade Department) to Robert J. 



Linn (Deputy Director, Physical Facilities, World Trade Department) (See Appendix A of NCSTAR 1-1) 
NIST NCSTAR 1-7, WTC Investigation 43 



Chapter 2 



Building codes largely relate required egress capacity to the size of the area served and the nature of the 
use of the area served. WTC 1 and WTC 2 were square buildings (roughly 207 ft by 207 ft, measured 
internally), with a gross square footage for each tenant floor approximately 42,850 ft 2 (3,990 m 2 ). The 
floor areas in the towers were typically one of two use categories: business or assembly. The distinction 
is important for calculating egress requirements as the number of people allowed in a given space would 
be significantly fewer if the space is used for office (business) activities, than if the same space were used 
for assembly activities (such as a restaurant or meeting space). The width and number of stairwells are 
then specified to equal or exceed the number of occupants on a floor. 

The size of the 'core' varied significantly throughout WTC 1 and WTC 2. Note that the size of the 
structural core (as defined by the location of interior load-bearing columns) did not change significantly 
from floor to floor. As discussed earlier in Chapter 2.2.3, however, on any given floor, the core space 
used for local elevators was reclaimed for leasable office space on successively higher floors within a 
zone. For example, while floors 42 to 48 had a core area of approximately 12,000 ft 2 (1,100 m 2 ), 
floor 105 had a core area of 6,800 ft 2 (630 m 2 ), or 57 percent of the core area of floors 42 through 48. 

The size of the core was important because some building codes (including the NYC Building Code) 
calculate occupant load on a net basis rather than a gross basis. A net basis reduces the square footage of 
a floor by an amount equal to the unoccupied space on a floor, such as elevator or machinery shafts, and 
common areas such as hallways. The logic of using net as a calculation basis rather than gross is that 
there does not need to be egress capacity provided for floor area where no occupants would be located. 
As an example, while floors in WTC 1 and WTC 2 had a gross square footage of approximately 
42,850 ft 2 (3,990 m 2 ), floor 105 had a net square footage of approximately 36,500 ft 2 (3,400 m 2 ), which 
yielded an occupant load of 365 persons. Floors 106 and 107 in WTC 1 were designated an assembly 
space for Windows on the World and were discussed previously in Chapter 2.2.1. By comparison, the 
occupant load on floors 42 - 48 would have been approximately 313 when calculated using 100 persons 
per ft 2 net. Generally, however, the floor with the largest occupant load dictates the overall design of an 
egress system. 

In addition to local changes in the size of the core space, the stairwells in the WTC towers changed floor 
location throughout the building, as well. This meant that the remoteness (or the distance the stairwells 
are located apart from one another) of stairwells varied, as well. The greatest separation distance between 
any two of the three stairwells, as measured by a walking path measurement (assuming that the building is 
fully sprinklered, which WTC 1 and WTC 2 were) determines the stairwell remoteness. At the two 
extremes of remoteness found in WTC 1 or WTC 2, floors 83 and higher had Stairwell A and Stairwell B 
located about 70 ft (21 m) apart, while on floors 77 - 82, Stairwell A and Stairwell C were located 
approximately 175 ft - 200 ft (54 m - 63 m) apart (depending upon the walking path on a particular 
floor). Coincidentally, WTC 1 was most heavily damaged on floors in the 90s (where the stairwells were 
the closest together) and all three stairwells were destroyed, while WTC 2 was attacked in a region where 
the stairwells were the most remote (floors 78 through 82) and one stairwell remained passable. The 
angle of the airplane impact, the length-wise orientation of the core, and the presence of elevator 
machinery near the passable stairwell may also have been contributing factors to the stairwell 
survivability, however. For context, most current codes require that two exits be located a distance apart 
no less than one-third of the diagonal distance of the area served (if the floor has full sprinkler protection) 
or no less than one-half of the diagonal distance of the area served (if the floor is not fully covered by 
sprinkler protection). One-half of the diagonal distance of the area served was 147 ft (45 m) and one- 



44 NIST NCSTAR 1-7, WTC Investigation 



Design of the World Trade Center Egress System 



third of the diagonal distance was 98 ft (30 m). Thus, in separate areas within the same building, stairwell 
remoteness distances would have been less than that required for sprinklered buildings, as well as greater 
than that required for unsprinklered buildings. 

As described earlier in Chapter 2.2, WTC 1 and WTC 2 had three stairwells, two 44 in. wide and one 
56 in. wide. The 44 in. stairwells were served by doors on each floor measuring 34 in. (1 m), while the 
56 in. stairwell was served on each floor by a door measuring 44 in. The NYC Building Code was 
selected due to the PANYNJ instruction to architects and engineers to adhere to the NYC Building Code. 
IBC and NFPA 5000 and NFPA 101 were selected because they are national model codes. See NIST 
NCSTAR 1-1 for further discussion of building codes. 

2.5.2 New York City Building Code (1968) 

Table 6-2 in §C26-601 required 100 ft 2 per occupant (net) for business occupancies, yielding a nominal 
occupant floor load of 365 persons per floor, based upon the largest net square footage (36,500 ft 2 ). In 
order to provide sufficient capacity for 365 persons, six and one-half units of exit width (at 60 people per 
22 in. unit) would have been required, yielding an allowable floor load of 390 people for business 
occupancies. A minimum of two stairwells would have been required for an occupant load less than 
500 people [§C26-602 Exits from Floors], each equally sized, as no more than 50 percent of the 
occupants can be served by a single exit. Two equally sized stairwells would have been 78 in. wide each. 
Three stairwells, two 44 in. wide and one 56 in. wide, would also provide the minimum egress capacity 
for business occupancy floors. 

Floors 106 and 107 in WTC 1 and floor 107 in WTC 2, having had occupant loads of over 1,000 persons 
each, would have required four stairwells to serve each floor. Thus, the 1968 NYC Building Code would 
have required theses spaces be served by a minimum of four stairwells (as the occupant load was greater 
than 1,000 persons). 30 The number of stairwells was not allowed to decrease in the direction egress 
travel, therefore, the entirety of WTC 1 and WTC 2 was required to have four stairwells. In 1995, once 
these spaces were considered formally between PANYNJ and NYC DOB, three stairwells were shown to 
provide adequate total capacity of 1,170 using the NYC Building Code exit reduction clause in §C27-367 
(Fasullo 1995). There was no mention, however, of the requirement for a fourth stairwell in either tower 
(which existed and continues to exist as a requirement in §C27-366), nor whether the floor and ceiling 
system satisfied the area of refuge requirement for a 2 h fire rating. 

Each stairwell would be required to have a door at least 0.9 m (36 in.) wide. The sum of two risers and 
one tread depth was required to be not less than 0.61 m (24 in.) nor greater than 0.65 m (25.5 in.). 31 
(Thus, the 'standard' 7 in. riser and 1 1 in. tread depth would satisfy this formula, at 25 in. [7 in. + 7 in. + 
1 1 in.]). According to Table 6-4 in the NYC Building Code, however, the stair rise, may not exceed 



Egress calculation performed in 1995 (contained as figures in Chapter 2.2.1) show a calculated egress capactity of 

1,170 persons for both floors 106 and 107 in WTC 1 and a calculated occupant load of 1,013 and 1,030 persons for floor 106 

and 107, respectively. Note that there was no mention in this memo of a fourth stairwell. 

From the NFPA Life Safety Code Handbook (2003 Edition), this note was made about the formula formerly used by NFPA 

and currently used by NYC: "This requirement was deleted because it was based on a 300-year-old French formula in which 

an inch was a slightly larger unit of measure than it is today. Moreover, people's feet and stride length - the basis of the 

formula - were somewhat smaller at that time. Also, the requirement was originally intended only for stairs of moderate 

steepness or pitch." 

NIST NCSTAR 1-7, WTC Investigation 45 



Chapter 2 



0.2 m (7.75 in.) and the tread depth must be greater than 0.24 m (9.5 in.). In 1968, the location of a floor 
exit was required to be as remote from the others as practicable. 

2.5.3 New York City Building Code (October 2003) 

As it pertains to the narrow scope of this egress analysis, the requirements related to the egress system in 
2003 would be identical to the requirements of 1968, with one significant exception: stairwell remoteness. 
In 1968, the location of a floor exit was required to be as remote from the others as practicable. New 
York City Local Law (LL) 16 (1984) imposed a remoteness requirement (not retroactive to an exiting 
building such as WTC 1 or WTC 2) of 9 m (30 ft) or one-third the maximum travel distance of the floor 
(55 m [180 ft]), whichever is greater, which for WTC 1 and WTC 2 was 55 m (180 ft). Thus, all floors of 
WTC 1 and WTC 2 had stairwell separations that exceeded the minimum separation distance requirement 
of New York City LL 16 (1984). 

2.5.4 International Building Code (2000) 

Chapter 10 of the 2000 International Building Code (IBC) require 100 ft 2 per occupant (gross), yielding a 
nominal occupant floor load of 429 persons per floor. A minimum of two stairwells would have been 
required (for occupant load less than 500 persons [Table 1005.2.1]), each equally sized. As WTC 1 and 
WTC 2 were fully sprinklered, Table 1003.2.3 requires a minimum of 0.005 m per occupant (0.2 in. per 
occupant) totaling 2.2 m (87 in.) of total stairwell width, or two 1.1m (44 in.) stairwells. 

Floors 106 and 107 in WTC 1 and floor 107 in WTC 2, having occupant loads of over 1,000 persons 
each, would require four stairwells to serve each floor. The four stairwells would be required to be 
maintained to grade, as the number of stairwells shall not decrease in the direction of egress travel. 
Additionally, the floor system would be required to have at least a 2 h fire rating. If two areas of refuge 
were built on floors 106 and 107 (each area holding at least one stairwell), the IBC would permit four 
44 in. stairwells. 

Section 1003.3.1 requires that each stairwell have a door at least 0.8 m (32 in.) wide. Section 1004.2.2.1 
requires that for fully sprinklered buildings, the stairwell doors be located a distance of no less that one- 
third the length of the maximum overall diagonal dimension of the building or area to be served (30 m 
[98 ft] for WTC 1 and WTC 2). This requirement was met on some floors, but not all floors, as discussed 
previously. 

2.5.5 NFPA 5000 and NFPA 101 - Life Safety Code (2003) 

The gross square footage for each WTC tower was 42,850 ft 2 (3,990 m 2 ). Table 7.3.1.2 requires 100 ft 2 
per occupant (gross) for a business occupancy, yielding a nominal occupant floor load of 429 persons per 
floor. A minimum of two stairwells would be required (for occupant load less than 500 persons 
[Section 7.4]), each equally sized. Table 7.3.3.1 required a minimum of 0.0076 m per occupant (0. 3 in. 
per occupant) totaling 3.3 m (129 in.) of total stairwell width, which may be satisfied by two 1.7 m 
(65 in.) stairwells, or three stairwells, sized at 1.1 m (44 in.) each. Section 7.2.1.2 requires that each 
stairwell have a door at least 0.8 m (32 in.) wide. 



46 NIST NCSTAR 1-7, WTC Investigation 



Design of the World Trade Center Egress System 



Floors 106 and 107 in WTC 1 and floor 107 in WTC 2, having occupant loads of over 1,000 persons 
each, would require four stairwells to serve each floor (Section 7.4.1.2). Four stairwells would be 
maintained to grade, as the number of stairwells shall not decrease in the direction of egress travel. 

Section 7.5.1.3 addresses the remoteness of stairwells, with the purpose "to minimize the possibility that 
more than one has the potential to be blocked by any one fire or other emergency condition." 
Additionally, Section 7.5.1.3.3 requires that for fully sprinklered buildings, the stairwell doors be located 
a distance of no less that one-third the length of the maximum overall diagonal dimension of the building 
or area to be served (30 m (98 ft) for WTC 1 and WTC 2). This requirement was met on some floors, but 
not all floors, as discussed previously. 



2.5.6 Comparison of Current Code Requirements 

Differences in Stairwell Occupant Capacity 

The IBC allows a reduction in egress capacity for fully-sprinklered buildings, to 0.005 m (0.2 in.) per 
person. Thus, while IBC would require two 44 in. stairwells for 429 occupants, NFPA Life Safety Code 
would require two 65 in. stairwells from tenant floors in of dimension similar to WTC 1 or WTC 2. 
While NYC Building Code also uses 0. 3 in. per person of required exit width, the calculation is net 
square feet, rather than gross square feet, effectively reducing the requirements. However, minimum half- 
units of exit width (12 in.) used in the NYC Building Code often force the designer to 'round up' the 
calculated egress capacity (from 365 to 390, e.g.). Two stairwells, each 78 in. in width, would be a 
minimum allowed by the NYC Building Code, as an alternative to the three stairwells (two 44 in., one 
56 in.). Table 2-3 summarizes the results of the calculation of the minimum number of stairwells for an 
office occupancy with a service area of 42,850 ft 2 gross (36,500 net) using each of the three building 
codes described above. 

Table 2-3. Minimum stairwell design for 42,850 ft 2 office plan. 



Building Code 


International Building 
Code (2003) 


NFPA Life Safety Code 
(2003) 


New York City Building 
Code (2003) 


Number and Width of 
Stairwells for a 
42,850 ft 2 Office Plan 


Two Stairwells 
44 in. each 


Two Stairwells 
65 in. each 


Two Stairwells 
78 in. each 



Net vs. Gross Occupant Load Calculations 

The IBC and NFPA model codes both calculate the number of occupants per floor (business occupancy) 
based on the gross square footage of the floor divided by 100. NYC Building Code, on the other hand, 
calculates the maximum occupant load by subtracting from the gross square footage, shafts, storage 
rooms, and stairs. Thus, the WTC had a design occupant load of 390 persons per floor, whereas IBC and 
NFPA 101 would have required egress capacity for 428 persons per floor. On the other hand, the egress 
system would have to reflect the higher occupant load, as well. WTC 1 and WTC 2 would have required 
7.5 (rather than 6.5) units of exit width if the occupant load calculation was on a gross basis rather than a 
net basis. Thus, three stairwells would need to have been 1.4 m (56 in.), rather than only Stairwell B, to 
accommodate the higher occupant load. 



NISTNCSTAR 1-7, WTC Investigation 



47 



Chapter 2 



Areas of Refuge and Egress Capacity 

When two stairwells are each located in separated areas of refuge, the capacity of each stairwell may be 
doubled. NYC Building Code, IBC, and NFPA 101 each permit doubling of a stairwell's capacity using 
this method. As discussed in Section 2.2.1, however, NYC Building Code also allows the capacity to be 
tripled when three stairwells are each separated from the other two by fire-rated partitions complying with 
requirements for areas of refuge. IBC and NFPA 101 do not allow tripling of stairwell capacity. 

Stairwell Remoteness 

NYC Building Code calculates the minimum separation distance for stairwells based upon a fraction of 
the longest travel distance on a particular floor (one-third, if fully sprinklered and one-half if not fully 
sprinklered). IBC and NFPA 101 calculate the minimum separation distance for stairwells based upon a 
fraction of the maximum diagonal of the floor or area served (one-third, if fully sprinklered and one-half 
if not fully sprinklered). 



48 NIST NCSTAR 1-7, WTC Investigation 



Chapter 3 

First-Person Data Collection and Analysis Methods 



3.1 PURPOSE AND SCOPE OF COLLECTION OF FIRST-PERSON DATA 

The purpose of first-person data collection was to capture the full range of occupant experiences from 
World Trade Center (WTC) 1, 2, and 7 from 8:46:30 a.m., when WTC 1 was attacked, until all survivors 
had successfully evacuated. The goal was to capture both common (frequent) evacuation experiences and 
unique observations or actions that may have contributed to a greater understanding of the events of 
September 11, 2001. Potential respondents included all occupants inside WTC 1 or WTC 2 between 
8:46:30 a.m. and 10:28:22 a.m. (when WTC 1 collapsed), building personnel, emergency responders, and 
family members who spoke to occupants inside WTC 1 or WTC 2 during the attack. 

3.2 METHODS 

To best capture both the generic evacuation experience and the unique observations and experiences, 
multiple interview methods were selected: face-to-face interviews, telephone interviews of a statistically 
representative sample of people, and focus group interviews. Each method contributed a unique strength 
to the overall objectives, complemented and contributed to understanding the data collected through the 
other methods, and established multiple measures of a variety of phenomena. In addition to interviews, 
published media accounts, video, and photographs were collected and analyzed. Each method is 
discussed below. 

3.2.1 Published First-Person Accounts 

National Institute of Standards and Technology (NIST) contracted with the National Fire Protection 
Association (NFPA) to collect first-person accounts from newspapers, radio and television programs, 
e-mail exchanges, and a variety of websites and to distill them into a searchable database (Fahy and 
Proulx 2003). Over a period of 18 months, a total of 745 first-person accounts were collected. These 
accounts had been published up to 14 months after the event. Although media accounts do not provide 
the rigor of a proper scientific study, they do present important insights into events. The objective of the 
analysis of the first-person accounts was to gain insight into the variability of human behavior and 
response time displayed during the evacuation, and to use the findings as a guide for additional 
investigation. For the NIST investigation, the accounts provided background for development of the 
telephone survey instrument and aided in identification of individuals with particularly compelling stories 
that were of interest for face-to-face interviews conducted as part of the investigation. 

A coding tool was developed for content analysis of the first-person accounts. Data were then entered 
into the database. The coding tool had 33 questions such as: "On what floor was the person?," "What was 
the first cue of the event?," "Was the person injured?," and "What were the conditions in the stairs?" Not 
every account provided answers for all 33 questions, since some accounts lacked certain details, but this is 
not unlike the situation of a respondent who did not answer some questions in a survey. Once the 
745 first-person accounts were summarized, multiple accounts from the same person were merged into 

NIST NCSTAR 1-7, WTC Investigation 49 



Chapter 3 



one, which provided accounts for 465 distinct individuals. (Some survivors provided multiple accounts 
through different sources.) Before any analysis began, the database was further limited to the 
435 building occupants who were actually in WTC 1 or WTC 2 on September 11, 2001. The accounts 
analyzed were from 435 individuals - 251 occupants of WTC 1 and 184 occupants of WTC 2 - 
representing occupants from low, middle, and high regions of both WTC 1 and WTC 2. 

The content analysis of first-person accounts has significant limitations. First, the actual questions asked 
by the journalists reporting the accounts are not usually contained in the accounts. Second, some details 
may have been left unreported; and third, more dramatic stories may have been over-represented. 
Consequently, while the results of the published accounts analysis cannot be generalized to the overall 
population of the towers of the World Trade Center, they provided valuable input to the NIST 
Investigation. 

3.2.2 Telephone Surveys 

The survey objective of the telephone interview phase of this study called for collecting 800 computer 
assisted telephone interviews (CATI) of persons occupying either of the WTC towers (WTC 1 and 
WTC 2) at the time of the terrorist attacks on September 11, 2001. The sample size of 800 and allocation 
of n=400 to each tower were chosen to maximize the statistical precision of estimates and projections 
within each tower. 32 Primary statistical analyses were in the form of tabulations and linear statistics 
(e.g., reporting of percentages and average/means). Estimates of percentages from tower-specific survey 
data (at n=400) exhibit sampling errors no greater than 2.5 percentage points, and 95 percent confidence 
intervals of percentages are no greater than ± 5 percentage points. This level of precision was more than 
adequate for examining issues of interest in this investigation. Within WTC 1 and WTC 2, independent 
proportionate stratified samples of survivors were drawn. In other words, each occupant of a particular 
tower had an equal probability of being selected. 

Population and Sampling Frame 

The total population of people eligible to participate in a telephone interview consisted of individuals who 
were inside WTC 1 or WTC 2 between 8:46:30 a.m. and 10:28:22 a.m. on September 11, 2001, with the 
exception of emergency responders (FDNY, NYPD, OEM, FBI. Secret Service, ATF, and others). The 
sampling frame (i.e., the list from which the sample was drawn) consisted of the names of occupants from 
badge lists for persons authorized to be present in WTC 1 and WTC 2 and was assumed to represent the 
entire population of individuals eligible to participate in telephone interviews. All occupants who worked 
or regularly visited the World Trade Center were required to provide personal data to PANYNJ in order 
to be issued a badge to clear through the security station at the entrance of each tower. The badge lists 
were provided to NIST by the Port Authority of New York and New Jersey. The lists provide name, floor 



32 Multivariate modeling such as correlation analyses, multiple linear regressions, and path analyses, are also a prominent part of 
the survey analyses. Like the tabulations, these analyses were conducted independently by tower. A sample size of n=400 per 
tower provides more than ample statistical power for the F tests used to determine the significance of the regression models 
(i.e., testing the null hypothesis that the ratio of explained variance to error/residual variance is equal to zero). For instance, in 
a multiple regression analysis featuring 20 independent variables, the sample size of 400, and 0.05 level of significance (Type I 
error), the power of the F test to detect an r statistic (i.e., proportion of explained variance) of 0.06 is just over 81%. See also 
Chapter 9 of Cohen, J., 1988, Statistical Power Analysis for the Behavioral Science, Lawrence Erlbaum Associates, Inc., 
Hillsdale, N.J. 

50 NIST NCSTAR 1-7, WTC Investigation 



First-Person Data Collection and Analysis Methods 



of occupancy, employer, and social security number (the only available means of uniquely identifying 
individuals). 

Representativeness of Badge List 

Confidence in the assumption that the badge list accurately represented the WTC population was 
increased by comparing independent 'lists' of occupants to the badge list: survivors who were interviewed 
by the media and lists of decedents. The three sources of data that were compared were (1) the list of 
decedents from CNN web site, (2) a media list of survivors, and (3) the badge list of occupants. 

One limitation is that the independent list of media interviewees may not have sufficient information to 
indicate whether the listed person should have had a badge (and thus been listed on the badge list). 

After comparing the media list with the badge list, it was determined that 134 (93 percent) individuals of 
144 selected from the media list were authorized to be at WTC 1 or WTC 2 the day of the tragedy. 
Approximately 2 percent of all individuals were definitely not supposed to be at WTC 1 or WTC 2 the 
day of the tragedy and insufficient information existed to determine positively if the remaining 5% were 
supposed to be at WTC 1 or WTC 2 on September 1 1, 2001. 

Based on Table 3-1, a conservative estimate of coverage was obtained by taking the number of persons 
from the media list who were definitely authorized to be at WTC (134), and comparing them to the total 
possible number of authorized individuals from that list -(134 + 7) = 141. The resulting conservative 
coverage rate of the badge list, estimated based upon media interviews with survivors, was (134/141), or 
95 percent. 



Table 3-1. Comparison of media interviewees and 


Dadge list. 


Status 


Frequency 


Percent 


Definitely authorized to be at WTC 1 or WTC 2 


134 


93% 


Definitely not authorized to be at WTC 1 or WTC 2 


3 


2% 


Not enough information to determine 


7 


5% 


Total 


144 


100% 



A similar analysis was conducted using the victim list published on the web site of CNN. As Table 3-2 
shows, a conservative estimate of coverage was obtained by taking the number of persons from the 
decedent list who were definitely authorized to be at WTC (2,141), and comparing that number to the 
total possible number of authorized individuals from that list (2,141 + 79 = 2,220). The resulting 
conservative coverage rate of the badge list was, as estimated from the CNN victim list, was 
(2,141/2,220), or 96.4 percent. 

Thus, the assumption that the badge list was a complete universe of possible WTC survivors from which 
to select a representative sample was determined to be valid. 



NIST NCSTAR 1-7, WTC Investigation 5 1 



Chapter 3 



Table 3-2. Comparison of CNN victim list and badge list. 



WTC Status 


Badge List Status 


Appears in Badge List 


Does Not Appear 
in the List 


Total 


Person authorized to be at WTC 


2,141 





2,141 


Not authorized to be at WTC 


N/A 


408 


408 


Insufficient information to 
determine WTC authorization 


N/A 


79 


79 


Total 


2,141 


487 


2,628 



Telephone Interview Sample Selection 

The badge list contained September 11, 2001, occupants, occupants who were absent on the day of the 
attacks, decedents, former occupants, and non-person listings (false names used in sample testing input by 
PANYNJ prior to delivery to NIST but not removed). This meant that a screening effort was needed to 
identify "eligible" badge list members - namely, those who were inside WTC 1 or WTC 2 during the 
attacks and survived. Moreover, the absence of telephone numbers for the badge holders on the list 
necessitated a tracking/locating effort. The primary tracking mechanism was to search public databases 
using commercially available batch matching and web-based search utilities. Consequently, a large 
sample was needed to generate the 800 completed interviews. 

The number of occupant selections drawn into the sample was contingent on four key design parameters: 

• The percentage of individuals from badge listings for whom a working telephone number could 
be found (initial estimate: 80 percent tracking success) 

• The percentage of badge listings that corresponded to a surviving WTC 1 or WTC 2 occupant on 
September 11, 2001 (initial estimate: 14 percent) 

• The cooperation rate for screening the occupants (initial estimate: 65 percent) 

• The interview response rate among September 11, 2001 survivors (initial estimate: 50 percent). 

In planning the CATI survey, a number of design parameters needed to be quantified in order to 
determine the number of persons to draw from the badge list. The expected disposition of the sample was 
developed using the parameters defined in the preceding paragraph. A total sample of 22,735 persons 
from the badge list was needed to generate the desired 800 completed interviews. The expected 
disposition by tracking efforts, screening and interviewing are discussed later. 

A reserve sample of about 14 percent (or about N=3,265) was added in the event additional respondents 
were needed due to unanticipated circumstance (if the eligibility rate was actually lower than anticipated). 
This brought the total sample size to 26,000. The reserve was initially held "in reserve" while the main 
sample was worked. Working the main sample allowed preliminary estimates of all design parameters to 
be monitored so that an informed decision could be made on the necessity of releasing none, some, or all 
of the reserve. 



52 



NIST NCSTAR 1-7, WTC Investigation 



First-Person Data Collection and Analysis Methods 



The badge list contained different counts of persons from each tower (slightly over 50,000 names for each 
tower), yet the sample design called for equal samples to be drawn from the collections of badge holders 
in WTC 1 and WTC 2. Thus, a disproportionate design (across tower strata) was employed. Within each 
tower, independent proportionate samples were drawn using stratification by floor (within tower), 
employer (within floor) and last name (within employer). This served to increase the statistical precision 
of the tower-specific samples. 

Thus, equal-sized samples of 13,000 selections were drawn from each of WTC 1 and WTC 2 badge lists. 
Each tower-specific sample was partitioned into 20 random replicates (comprising 5 percent of the total), 
and the reserve sample was determined by the last several random replicates for each tower. It is 
important to note that all badge holders from WTC 1 floors 92 and above were omitted from sampling 
because there were no survivors from those floors. 

Table 3-3 summarizes the final disposition of the CATI sample and the total (locating) sample. The table 
is comprised of two sets of rows. The top set pertains to the CATI sample and represents those sample 
persons for whom an initial telephone number was identified prior to commencing the CATI survey 
operations. The bottom set of rows with the heading "Total Sample Disposition" represents the results of 
the locating/tracking effort used to identify usable telephone numbers associated with the sample subjects. 
(Recall that only name, SSN, and employer were available; no other contact information was readily 
available.) 

The bottom set of rows shows that telephone numbers were identified for just over three quarters 
(76.7 percent) of the sampled subjects. Moreover, this rate was fairly uniform across towers. The 
19,923 individuals with an initial telephone number were then loaded into the CATI sample management 
system for calling. Ultimately, all reserve respondents were used in the telephone survey. In the initial 
design parameters, it was assumed that 82 percent of the subjects would be locatable. While 76.7 percent 
is close, many of the numbers were obsolete (e.g., disconnect, wrong number) and necessitated additional 
tracking during CATI operations. Ultimately, by the end of data collection, only half the sample 
(49.5 percent) represented confirmed contacts with respondents. 

The top set of rows in Table 3-3 presents the final disposition of the sample by tower as well as for the 
overall sample. Several statistics in the percentage distribution (rightmost) column are notable. First, 
NuStats (under contract to NIST) was unable to contact subjects for half the sample (50.5 percent), due to 
failures to answer the phone, answering machines, unusable numbers (e.g., wrong number, disconnected, 
business), etc. Most of these unusable telephone numbers represent "unbeatable" subjects - subjects for 
whom the initial telephone number was incorrect. It bears reiterating that substantial additional attempts 
to locate individuals during CATI operations were conducted using powerful subscription-based web- 
based search engines. Unfortunately, little information was available for these individuals. 



NIST NCSTAR 1-7, WTC Investigation 53 



Chapter 3 



Table 3-3. Disposition ol 


the CATI sample and the total sample by tower. 


CATI Disposition: 


WTCl" 


WTC 2 a 


Total 


% Distn 


Interview 


427 


376 


803 


4.0 % 


Partial Interview 


47 


37 


84 


0.4 % 


9/11 decedent 


20 


40 


60 


0.3 % 


Other decedent 


49 


39 


88 


0.4 % 


Not Eligible 


3,712 


3,752 


7,464 


37.5 % 


Language Barrier 


135 


129 


264 


1.3% 


Eligible Refused to Interview 


138 


139 


277 


1.4% 


Other Refusal 


224 


181 


405 


2.0 % 


Respondent not Interviewed 


247 


168 


415 


2.1% 


Can't contact/locate Respondent 


4,987 


5,076 


10,063 


50.5 % 


CATI TOTAL 


9,986 


9,937 


19,923 


100.0 % 


Total Sample Disposition: 


WTCl 


WTC 2 


Total 


% Distn 


Found initial telephone # 


9,986 


9,937 


19,923 


76.6 % 


Unable to find a telephone # 


3,014 


3,063 


6,077 


23.4% 


SAMPLE TOTAL 


13,000 


13,000 


26,000 


100 % 



a. Table data are unweighted. Tower location as indicated in the badge list and may differ from reported tower location. 

Second, the badge list contained a number of ineligible subjects (37.5 percent) - individuals on the badge 
list but not in the building on the morning of September 11,2001. An assessment of eligibility rates 
appears later. Third, the badge list included decedent names (0.4 percent) - some from the 
September 11, 2001, attack (0.3 percent) and others from causes not necessarily related to September 11, 
2001 (e.g., cause unknown, natural causes, 0.1 percent). Most of the September 11, 2001, decedent 
names were encountered due to a difference between the full (formal) name of the subject and the name 
that appeared on the badge list (e.g., the badge list sometimes contained maiden names, middle names, 
nicknames, misspelled first or last names, out of sequence names, titles, and so on). This impeded the 
ability to remove known decedent names prior to calling. 

The final outcome rates of the CATI operations are presented by tower in Table 3-4. The table shows 
screening rates, interview rates, and rates of eligible occupants (among those who responded to the 
screening questions). The first row shows that screening response rates were relatively uniform across 
the towers, at about 46 percent. In other words, approximately 46 percent of successful telephone 
contacts resulted in determining whether the potential respondent was present at WTC 1 or WTC 2 on 
September 11, 2001. Similarly, interview response rates (among screened eligible subjects) were 
relatively stable across towers, at about 49 percent. 

The eligibility rates were higher than expected - about 18 percent overall compared to the 14 percent 
expected. The eligibility rate among WTC 1 subjects was slightly higher than those of WTC 2. 
However, the overall response rates are essentially uniform across towers, at 22.6 percent. 



54 



NIST NCSTAR 1-7, WTC Investigation 



First-Person Data Collection and Analysis Methods 



Table 3- 


-4. Summary disposition rates by tower. 




Disposition Rate" 


WTC 1 


WTC 2 


Total 


Screen 


46.5 % 


45.8 % 


46.1 % 


Interview 


48.6 % 


49.5 % 


49.0 % 


Eligibility 


18.9% 


16.7% 


17.8% 


Overall 


22.6 % 


22.7 % 


22.6 % 



a. Definitions for "Rates" consistent with American Association of Public Opinion Research (AAPOR) Standards, 
which may be found at http://www.aapor.org/pdfs/standarddefs_3. 1 .pdf . 

The telephone interview protocol resulted in 803 interviews with individuals who evacuated WTC 1 or 
WTC 2 on September 11, 2001, after 8:46:30 a.m. These 803 individuals were interviewed in roughly 
equal proportion (N = 440 for WTC 1 and N = 363 for WTC 2) between the two buildings. The interview 
results can be generalized to the entire population of survivors in both buildings with a high degree of 
statistical confidence. 

Telephone Questionnaire 

The telephone interview was conducted by trained interviewers using a computer program that provides 
questions and answer categories for the interviewer. Prior to being contacted by telephone, subjects 
received a letter that outlined the scope and purpose of the investigation and the purpose of the interview, 
and indicated that a telephone call would come several days later. A full informed consent statement also 
appeared in the letter, as well as in the script for the calls. 

When interviewers reached the subjects by telephone, they described the survey, the confidentiality of 
responses, the length of the interview, and the voluntary nature of participation. Subjects were then asked 
if they wished to participate, which served as the means of obtaining oral informed consent. 

The telephone interview instrument (see Appendix A in this report for the complete instrument) included 
the questions, variable names, response options, and skip patterns taken directly from the computer 
program used by the interviewers. Variable names are used as shorthand for subsequent data analysis. 
Questions had a variety of response option categories: multiple choice, interval, Likert scale, or open- 
ended. Open-ended responses were minimized where possible due to the analysis burden and the fact that 
face-to-face interviews also were being conducted. Skip patterns reduced the burden on the respondent 
by skipping questions that would not apply to a particular respondent. For example, a respondent would 
not be further questioned about fire drills if he or she had not received fire drill training. Subsequent 
discussions of the questions will indicate whether a respondent was read a list of choices or was expected 
to give a free response. 

The interview, which typically lasted approximately 20 minutes, was designed around five primary 
groups of questions, covering emergency training and preparedness, three stages of evacuation 
experience, and background information about each respondent. 



NISTNCSTAR 1-7, WTC Investigation 



55 



Chapter 3 



Emergency Training and Preparedness 

The first group of questions measured the extent to which an occupant had any special level of knowledge 
about the building, other than what would be obtained by performing his or her job. The most prevalent 
special knowledge would be formal evacuation training, or fire drills. If respondents indicated that they 
participated in evacuation training during the 12 months prior to September 11, 2001, further questions 
were asked about the content of the training. The occupant's understanding of the emergency procedures, 
or the way it was 'supposed to go,' was also measured. Next, a Likert scale 33 measured the perceived 
usefulness of the evacuation training in the context of egress experiences on September 11, 2001, ranging 
from very helpful to very unhelpful. Finally, the respondent was asked whether he or she knew that there 
was a floor warden for his or her floor. 

Initial Experience on September 11, 2001 

The second group of questions covered the first moments of the September 11, 2001 attack on the World 
Trade Center as experienced by the respondent, also known as the initial awareness period. The manner 
in which a person first became aware that something was not normal (whether in the building or the 
neighboring building) may have influenced subsequent decisions. Examples of awareness channels may 
include sensory perception (such as feeling, hearing, or seeing the building shake; seeing or smelling fire 
or smoke) or may include a conversation with a person inside or outside the WTC complex. Next, the 
respondent was asked to provide context to the initial moment of awareness. Context was first created by 
identifying what activity the respondent was performing. Activities included, but were not limited to, 
working, conversing with coworker(s), eating, or participating in a meeting. The respondent was then 
asked to recall the number of other people he or she was with at the first moment of awareness. People in 
groups often defer to group decisions rather than making their own evacuation decisions. Next, a list of 
observations was read aloud, and the respondent was asked to indicate whether he or she noticed that 
event during the period of initial awareness. These events included smoke, fire, fireballs, collapsed walls, 
jet fuel, severely or fatally injured people, sprinklers going on, fire alarm sounding, power outage or 
flickering lights, fallen ceiling tiles, and extreme heat. The event proximity was probed for every 
affirmative response to determine whether the observed event was in the immediate area or outside the 
building. If no affirmative responses were indicated, the respondents were asked whether they observed 
any disaster related events not previously mentioned. Finally, the extent of any injuries to the respondent 
or those in the immediate area was ascertained, as well as whether the respondent felt that his or her life 
or the lives of other people were in danger. 

Interim Experience on September 11, 2001 

The format of the interim experience group of questions mirrored the format of the initial awareness 
questions. The interim time period was defined as the time after initial awareness, but before the person 
entered a stairwell or elevator to leave the building. This time period may have ranged from moments to 
tens of minutes. The objective of the interim period questions was to determine what motivated/forced 
people to either immediately evacuate or delay their evacuation by some period of time. 



A Likert scale measures the degree to which the respondent agrees or disagrees with a statement. In this case, the scale 
measured helpfulness, including very helpful, helpful, unhelpful, and very unhelpful. A neutral response was not included. 

56 NIST NCSTAR 1-7, WTC Investigation 



First-Person Data Collection and Analysis Methods 



Information about the nature of the event obtained during the initial period often forms the basis for 
decision-making during the interim period. For example, many people may have found the environmental 
cues from the initial awareness period sufficient to initiate an immediate evacuation. Others may have 
required additional information in order to feel comfortable leaving the workplace. Occupants could have 
obtained information in two ways: passively and actively. Passive information is information received 
without seeking it. In other words, the information is received regardless of whether the person feels it is 
needed. Active information is information which the respondent actively seeks and considers important 
with respect to his or her decision to evacuate. In the interview, respondents were first asked whether 
they received any additional information about the event during the interim period. If so, the source 
(who), the nature (what), and the channel (how) of the information was probed. Next, additional 
information sought by the respondent was probed, including the source, nature, channel, and whether the 
process was successful in gathering additional information. 

The perception of risk to the respondent's life, as well as the lives of others, was asked in the same way as 
during the initial period, to determine whether the sense of risk was increasing or decreasing over time. 
The interviewer probed about the activities of other people in the proximity of the respondent, which may 
have influenced the respondent's subsequent choices. Whether other people began evacuating prior to the 
respondent was specifically asked. Next, respondents were asked about the activities undertaken during 
the interim period, as well as activities that they wanted to carry out but could not. These activities 
included work-related actions, such as saving files or shutting machines down; personal actions, such as 
gathering belongings or calling people; and emergency-related actions, such as fighting fires/smoke, and 
searching for or helping others. If a respondent was unable to accomplish an action, the action and the 
reason for being prevented from doing so was captured. 

As with the initial period, any observations of building damage were collected. If the respondent received 
help in any way before initiating evacuation, the nature and source of the assistance was determined. The 
respondent was asked for the primary cue was that initiated his or her evacuation on September 11, 2001 
and how many minutes passed before initiating evacuation. Finally, respondents were asked whether 
anything prevented them from evacuating sooner than they reported. 

Evacuation Experience on September 11, 2001 

The next group of questions which followed the evacuation sequence to its completion, focused on time 
spent in the stairwell and/or elevator(s). Respondents were first asked whether they began their 
evacuation alone or with other people. Which stairwell (or elevator) the respondent entered was collected 
as either the stair identification letter (A, B, or C) or the geographic location, if known. Knowing where 
the stairwell emptied at the bottom could also narrow down which stairwell was used, which was 
collected near the end of this group of questions (Stairs A/C [44 in. wide] emptied out to the upper, 
Mezzanine Level, while Stair B [56 in. wide] went to the lower, Concourse Level). Next, the 
respondent's rationale for using a particular stairwell was probed. The respondent was then asked 
whether he or she left the stairwell or turned back for any reason during the evacuation and, if so, why. 

Some events and features of the stairwells aided the progress of the evacuation, while other features 
constrained the progress of the evacuation. The following features or events were identified to the 
respondents, who were asked to indicate whether it was an aid to their egress: instructions or assistance 
from their floor warden, a police officer, or firefighter; support/encouragement from others; exit signage; 



NIST NCSTAR 1-7, WTC Investigation 57 



Chapter 3 



and photoluminscent paint. The following items were identified to determine whether they constrained 
the evacuation: crowded stairwells, counterflow (people moving up the stairs, against the flow of 
occupants), disabled or injured people being taken down the stairwell, locked doors, poor lighting, 
confusing or missing signage, and lack of clear instructions. 

As with the initial and interim time periods, environmental cues related to fire, smoke, jet fuel and other 
disaster-related observations were probed, as well as whether the observation was in the immediate area 
or outside the tower. The final question about the respondent's own evacuation estimated the elapsed 
time from entering the stairwell until they left the building. A concluding evacuation question determined 
whether they knew why someone on their floor did not survive the WTC attack, if applicable. 

Respondent Background 

The final group of questions explored the background of the respondent relevant to evacuation. The first 
question identified any pre-existing disabilities or injuries which made evacuation more difficult. The 
respondent's age, gender, and primary language were collected. If the respondent was working in the 
building prior to 1993, they were asked whether they were present during the February 26, 1993 bombing. 
If so, respondents were asked questions about their evacuation experience on that day. 

The interview concluded with an open-ended opportunity for the respondent to say anything additional 
about their evacuation experience on September 11, 2001. Respondents who indicated that they had a 
disability, were near the floors of impact, observed fire, smoke, or fireballs in their immediate area, or had 
a role of building responsibility on September 11, 2001, were asked if they would be willing to participate 
in a follow-up face-to-face interview. 

Telephone Interview Response Rate Analysis 

The response rate analysis of the telephone interview sample indicated an inverse relationship between 
floor height and the rate of response in WTC 1, as shown in the last column of Table 3-5. In other words, 
an individual was somewhat less likely to complete a telephone interview if they were high in WTC 1 
than if they were lower in WTC 1 . The non-response weight adjustment is the inverse of the overall 
response rate. For example, the inverse of 25.3 percent is 3.95. 34 In general, the weight adjustment for 
WTC 1 indicates that representative results should reflect that a single interview with a respondent high in 
the building is representative of more occupants than a single interview with a person lower in the 
building. 



1 / 0.253 = 3.95. A lower overall percentage would, therefore, yield a higher weight adjustment. 
58 NIST NCSTAR 1 - 7, WTC Investigation 



First-Person Data Collection and Analysis Methods 



Table 3-5. Response rate analysis for WTC 1. 


Floor 
Stratum 


Number of 
Selections 


Number of 
Interviews 


Screen 


Eligibility 


Interview 


Overall 


Non-Response 

Weight 

Adjustment 


1 to 42 


4464 


256 


46.2 % 


22.6 % 


54.8 % 


25.3 % 


3.95 


43 to 75 


3714 


137 


48.6 % 


16.6% 


45.8 % 


22.3 % 


4.49 


76 to 91 


1802 


34 


42.7 % 


14.7 % 


30.1 % 


12.9% 


7.78 


Floor missing 


6 





50.0 % 


0.0 % 


NA 


NA 




Total 


9986 


427 


46.5 % 


18.9% 


48.6 % 


22.6 % 





Key: NA, not applicable. 

A similar analysis of telephone interview response rates for WTC 2 (shown below in Table 3-6) did not 
indicate a significant need to weight the results; however, to be consistent with WTC 1 analysis, the 
results were weighted. 





Table 3-6. Response 


rate analysis for WTC 2. 




Floor 
Stratum 


Number of 
Selections 


Number of 
Interviews 


Screen 


Eligibility 


Interview 


Overall 


Non-Response 

Weight 

Adjustment 


1 to 42 


4339 


143 


44.8 % 


14.8 % 


49.7 % 


22.3 % 


4.49 


43 to 75 


3187 


134 


45.0 % 


17.7% 


52.8 % 


23.8 % 


4.21 


76 to 110 


2203 


94 


48.3 % 


19.5% 


45.2 % 


21.8% 


4.58 


Floor missing 


208 


5 


50.5 % 


9.5 % 


50.0 % 


25.2 % 


3.96 


Total 


9937 


376 


45.8 % 


16.7% 


49.5 % 


22.7 % 





All subsequent telephone interview data analysis in this report reflects weighting of the results in order to 
more accurately generalize the results. By convention, when a sample number is indicated (n= ), the 
sample number will be the actual number of responses. Where percentages are indicated, however, the 
percentages were weighted to allow for generalization, unless otherwise indicated. 

The source of differential non-response for floors 76 to 91 in WTC 1 when compared to floors 1 to 75 
was not specifically identified. 



3.2.3 



Face-to-Face Interviews 



The objective of the face-to-face interview segment was to gather first-hand accounts and observations of 
the activities and events inside the buildings on the morning of September 11, 2001. Using this approach, 
NIST identified previously unknown information, evaluated technical hypotheses, and explored conscious 
motivations for occupant behaviors, while allowing for comparisons to the telephone interview data. 
There was no recording of the face-to-face interviews, other than random selections for quality control 
purposes. It is estimated that the average face-to-face interview lasted approximately two hours. 

The methodology for the face-to-face interviews is a synthesis of the Behavioral Sequence Interview 
Technique (BSIT), originally developed by Keating and Loftus (Keating and Loftus 1984), and the 



NISTNCSTAR 1-7, WTC Investigation 



59 



Chapter 3 



Cognitive Interviewing Method (CIM), originally developed by Fisher (Fisher et al. 2000) and Geiselman 
(Geiselman 1986). These two interviewing methodologies were developed with the purpose of assisting 
persons in retrieving more comprehensive and accurate memories of incidents and sharing important 
details. Both approaches begin by allowing the informant to retell an unimpeded account without 
interruption by the interviewer, and both initially employ a chronological retelling of information. 
However, BSIT was designed to yield a database of qualitative information that could be subjected to 
systematic analysis and consolidation, while CIM was designed to facilitate investigative interviews. 
Since the Investigation is pursuing both goals (i.e., creation of a database of evacuation-related behaviors 
and an investigatory capture of information relevant to outcomes), the methodology combines these two 
approaches. 

Cognitive interviewing has been the subject of many empirical investigations. Fisher et al. (Fisher, 
Brennan, and McCauley 2002) summarized these findings, demonstrating that the methodology 
significantly increases the amount of information recalled without affecting rate of errors. Interviewing a 
large number of informants will allow corroboration of information, thereby compensating for the likely 
increase in the absolute number of errors. Accordingly, it is likely that this approach will be productive in 
achieving a holistic view of the building evacuations. 

The face-to-face interview methodology involved occupants who may have observed (knowingly or 
unknowingly) events important to completion of the objectives of the investigation. 

Enumerating the population: The population included the entire occupant and building management 
population of World Trade Center WTC 1,2, and 7. 

Selecting the sample: The face-to-face interview sample was developed by identifying: (a) individuals 
who identified themselves as being willing to share information relevant to the objectives of the NIST 
investigation, (b) individuals identified from the telephone interview sample as having experiences or 
observation requiring further exploration, and (c) the snowball quota sample approach. A snowball quota 
sample approach asks individuals for the names of other people who may meet the selection criteria for 
the study. The people identified are subsequently contacted and asked the same question. The process 
continues until the quota has been reached. 

Data Collection: The face-to-face interviews followed a four step technique, including unimpeded, open- 
ended narrative, a structured narrative, technical probes, and closed-ended questions. Each step is 
described more fully below. 

Step 1: Unimpeded open-ended narrative account: Both BSIT and CIM begin the process by asking 
the participant to chronologically recount his or her "story." For the NIST investigation, the starting point 
was when it became apparent that something unusual had occurred on the morning of 
September 11, 2001. The ending point was when the participant felt that he or she reached a location 
where he or she felt safe (or, alternatively, when he or she successfully reached the exterior of the 
building). Researchers and practitioners involved with cognitive interviewing believe that starting the 
face-to-face interviews in this manner both improves recall and helps build rapport between the 
participant and the interviewer. Fisher et al. also noted that asking questions may interfere with recall 
because a participant must divide his or her mental resources between recall and listening to the 
interviewer's questions (Fisher, Falkner, Trevisan, and McCauley 2000). During the open-ended 



60 NIST NCSTAR 1-7, WTC Investigation 



First-Person Data Collection and Analysis Methods 



narrative account, the interviewer often records notable information that was used for the probing phase 
conducted later. 

Step 2: Structured narrative account: After participants completed their stories, interviewers prompted 
the respondents to go through the story again, but this time working cooperatively with the interviewer to 
record entries into a table. This approach was employed by BSIT for three primary reasons: (1) to yield a 
structured account that can be entered into a database without further processing; (2) to avoid the biasing 
effects of having interviewers ask specific questions; and, (3) to enhance the effort at recall put forward 
by participants by encouraging their active collaborative participation, an advantage to open-ended 
formats as noted by Fisher et al. (Fisher, Falkner, Trevisan, and McCauley 2000). 

Each row of the table represented a single action. The approach was used based on the hypothesis that 
people encode narrative memory in a manner consistent with this format, thus facilitating both recall and 
data entry. Each column of the table represented three essential components of actions: a cue, an action, 
and the reason for taking that action. Cues can be either external (e.g., signs of a fire, someone saying 
something) or internal (e.g., remembering another means of escape). Actions are expressed using specific 
action verbs (i.e., "ran" instead of "went") and may include objects (e.g., a fire extinguisher) used by the 
informant. Reasons are the intentional, goal-directed base for the action. The interviewer used the 
participant's own words to the greatest extent possible. Participants were asked to review the data for 
accuracy. 

Table 3-7 is an example of actions recorded in this manner. 

Table 3-7. Example tabular face-to-face interview data entry. 



Cue 


Action 


Reason 


I heard but couldn't see 
someone yell "I've found a 
clear path" 


So I stumbled in the dark towards where 
I thought the voice came 


So that I could find a way to 

escape 


My path was blocked by debris 


So I called out to whoever yelled, "I'm 
near the reception area. Where are 
you?" 


To try to get a better idea about 
where the person was 



Experimental findings in psychological research on memory (Nillson 2000) suggest that when people 
perform actions, their abilities to verbally recall those actions are significantly improved. Script theory 
(Schank and Abelson 1977) suggests that people naturally organize their knowledge of actions using 
narrative sequences of actions structured around their pursuit of goals. However, in the case of 
WTC interviews, gaps in the narrative are anticipated, especially given the long period of time that has 
elapsed between the event and the interview. For the NIST investigation, however, interviewers 
encouraged participants to report only those memories about which they were confident really occurred to 
them. 

Step 3: Probing for specific information: After completing the structured narrative account, 
interviewers asked specific open-ended questions (probes) intended to elicit specific information of 
particular value to the investigation. While some of this information was likely to have been part of the 
structured narrative account, participants could also recall other valuable information as well in response 
to probes. 



NISTNCSTAR 1-7, WTC Investigation 



61 



Chapter 3 



Interviewers may use "context reinstatement" from CIM to improve recall of important information, 
because laboratory experiments have demonstrated that contextual cues enhance recall of related 
information. Fisher et al. explain that context reinstatement may enhance recall because people use 
multisensory coding of events. Using this mnemonic method, interviewers ask participants to "mentally 
recreate the external environment, and their affective, physiological, cognitive, and emotional states that 
existed at the time of original event" (Fisher, Falkner, Trevisan, and McCauley 2000). 

Depending on the population, probes were used to try to elicit information including, but not limited to: 

• Location of the informant at the time of certain marker events (e.g., location in WTC 1 when 
WTC 2 collapsed); 

• Fire conditions (e.g., fire and smoke); 

• Other cues of interest (e.g., the smell of jet fuel); 

• Presence and activities of persons with disabilities; 

• Use of elevators by self or others; and 

• Knowledge of any obstacles to their progress while using the stairs. 

Because information about many of these areas of concern required precise responses, questions for open- 
ended probes were developed collaboratively between the contractor and NIST. Responses to probes 
were recorded using standardized formats where feasible. For example, all participants who observed 
smoke were asked to estimate the smoke density qualitatively. 

Quality Control for Face-to-Face Interviews: With the respondent's consent, some of the face-to-face 
interviews were audiotaped. The audio tapes were used for quality control of both data collection and 
interview quality. The audiotapes were periodically reviewed to ensure that interviewers precisely 
followed the protocol and conformed to administrative requirements. 

Interview Responses 

Over 200 face-to-face interviews were conducted with survivors from WTC 1 and 2; 131 interviews were 
performed with survivors of WTC 1, and 73 interviews with survivors from WTC 2, with the remainder 
from WTC 7 or with building personnel who moved between both WTC 1 and WTC 2. An attempt was 
made to interview occupants from each zone: low, middle, and high (at least 20 occupants were 
interviewed from each zone) of WTC 1 and WTC 2), as well as to interview occupants with unique 
experiences: occupants who witnessed fireballs, occupants with mobility impairments, occupants trapped 
in elevators, and occupants near or above the floors of impact. Occupants with special roles in the 
building, such as floor wardens and PANYNJ employees or contractors, were interviewed. Finally, six 
family members who spoke with occupants inside WTC 1 or WTC 2 after 8:46:30 a.m. on 
September 11, 2001 were interviewed. 



62 NIST NCSTAR 1-7, WTC Investigation 



First-Person Data Collection and Analysis Methods 



3.2.4 Focus Groups 

Williams reported that in a group setting, people provide cues that evoke memories in others and that 
social pressures mediate against reporting misrepresentations of what they recall (Williams 1990). The 
overall goal of the WTC focus group interviews was to elicit accurate group representations of specific 
events or themes. The six focus groups and the corresponding objectives were: 



• 



• 



• 



Occupants located near the floors of impact. The objective of conducting a focus group with 
people near the floors of impact was to obtain information on the extent of the building damage 
and how the damage influenced the evacuation process. 

Floor wardens. The objective of the floor warden focus group was to explore the implementation 
of the floor warden procedures and the effect those actions had on the evacuation of the occupants 
on a floor and the evacuation of the floor warden themselves. 

Mobility-impaired occupants. The objective of this focus group was to explore the effect of a 
disability on the evacuation of the occupant and any other individuals who may have assisted or 
otherwise been affected by the evacuee. 

Persons with Building Responsibilities: The objective of the focus group with persons with 
building responsibilities was to capture the unique perspective custodians, security, maintenance, 
or other building staff. 

Random evacuees in WTC 1. The objective of the focus group with random evacuees in WTC 1 
was to further explore the variables used in the causal modeling (i.e., those obtained from the 
hone interviews) that best explained evacuation delay and normalized stairwell evacuation time, 
including environmental cues, floor, and activities. 

• Random evacuees in WTC 2. The objective of the focus group with random evacuees in WTC 2 
was to further explore the variables used in the causal modeling (i.e., those obtained from the 
hone interviews) that best explained evacuation delay, including environmental cues, floor, risk 
perception, and use of elevators. 

Sample selection: The people selected for inclusion in this study were selected using non-probability 
sampling procedures, i.e. snowball quota sampling (Blalock 1972; Cochrane 1977). Respondents 
contacted for face-to-face interviews or for other reasons were asked to provide the names and contact 
information for people they knew in each of the categories described above. Every effort was made to 
include approximately five people in each of these categories in the focus group study. 

Data collection: The data collected in the focus group study produced qualitative and detailed narrative 
accounts of the experiences of several groups of people, including those near the floors of impact and 
those having mobility impairments. The focus group discussion was moderated by a trained and 
experienced facilitator. Two notetakers recorded the discussion and later compiled into a single summary 
of the focus group. 



NIST NCSTAR 1-7, WTC Investigation 63 



Chapter 3 



3.2.5 Audio, Video, Photographic, and Records Collection 

Numerous emergency communications were recorded between 8:46:30 a.m. and 10:28:22 a.m., a period 
during which building occupants, WTC personnel, emergency responders, and people outside the 
WTC complex used radios and telephones to cope with the unfolding disaster. Emergency 9-1-1 call 
records, made available by the City of New York, were analyzed in both audio and transcript format. 
Communications at the WTC complex, transcripts of which were made publicly available by PANYNJ, 
were analyzed in both audio and transcript format. These communications included radio channels 
internal to the WTC complex, such as maintenance, vertical transportation, security, and PAPD. They 
also included communications external to the WTC complex, such as NYPD, EMS, PATH, and Newark 
airport. 

In stark contrast to the number of photographs and video taken of the outside of WTC 1, WTC 2, and 
WTC 7, very few video or photographic records from inside WTC 1, WTC 2, or WTC 7 survived. A 
significant number of records related to the design and maintenance of the egress and communications 
systems, however, were collected and analyzed. In particular, records regarding the identities of the 
occupants authorized to enter the WTC complex (referred to as the badge list), architectural drawings, 
tenant alteration applications, building upgrades, emergency plans and procedures, and training materials 
contributed to a better understanding of the egress system performance on September 11. 

Finally, complaints filed with the Occupational Safety and Health Administration (OSHA) were 
reviewed. The issues raised by complainants, a combination of surviving occupants and families of 
victims, guided the development of interview instruments and identified additional avenues of 
investigatory pursuit for emergency preparedness and evacuation system performance. 

3.3 ANALYSIS OF FIRST-PERSON DATA 

A systematic, comprehensive approach was required in order to consolidate the enormous volume of data 
collected or made available. Two primary techniques were utilized in order to capture the full range of 
the data collection: quantitative and qualitative analysis. Neither technique would have sufficed by itself, 
as the two techniques were highly complementary. In general, the quantitative data analysis was used 
with the telephone interview data, while the qualitative data analysis was used with face-to-face 
interviews, focus groups, and emergency communications. A number of analysis techniques were 
considered, however, the approaches described below were selected to provide the highest quality results 
across the three interview methods. 

Quantitative Data Analysis 

The quantitative data analysis, based on to a telephone interview data set collected according to rigorous 
statistical standards, provided the ability to generalize findings and conclusions to the entire population of 
WTC 1 and WTC 2 survivors. The results of the 803 telephone interviews were archived and analyzed 
using SPSS 12.0.1, 35 a statistical analysis software package. This package provided the ability to apply 
weights to the data, compute mean, median, mode, skewness, and other relevant statistical measures, 
recode data using expert judgment, and automatically produce reports of subsets. 



See Federal Government product disclaimer contained at the beginning of this report. 
64 NIST NCSTAR 1 - 7, WTC Investigation 



First-Person Data Collection and Analysis Methods 



Qualitative Data Analysis 

The purpose of the face-to-face interviews and focus groups was to explore the events of 
September 11, 2001 at WTC 1, WTC 2, and WTC 7 in an investigatory manner, allowing the individuals 
to communicate their experiences, observations, and thoughts outside the constraints of a closed-ended 
interview format. While a framework of data collection format and a time frame (the morning of 
September 11, 2001) was imposed, respondents had complete freedom to express their experiences in 
their own words. The primary value of the qualitative data analysis was two-fold: first, to collect 
information on the entire range of observations and experiences, and second, to enhance and provide a 
deeper understanding of topics explored and generalized in the telephone interviews. ATLAS. ti 4.1 35 was 
used to conduct qualitative data analysis. Each face-to-face interview and focus group was coded for over 
130 different types of information, resulting in several thousand individual codings. The codings, along 
with a brief explanation of each coding, are included in Appendix B at the end of this report. The 
resulting dataset was queried for targeted exploration of various factors of interest. 

Protection of Human Subjects 

To ensure the protection of the respondents and to comply with the Common Rule for the Protection of 
Human Subjects, the protocols and informed consent forms for the conduct of the telephone interviews, 
face-to-face interviews, and focus groups were reviewed and approved by an institutional review board 
and by NIST. NIST NCSTAR 1-7B summarizes the procedures used in first-person interviews. 

Confidentiality of Respondents 

NIST noted the importance of maintaining the confidentiality of the respondents. Contractors and 
retained experts were bound by contractual obligation to protect the confidentiality of all interview 
respondents, whether interviewed by telephone, face-to-face, or focus groups. No identifying information 
(name, gender, floor number, job title, etc.) has been included in this report. Individuals may have 
previously spoken to the media or other individuals about their experiences and observations, however, 
which could diminish NIST's ability to protect an identity. Interviews conducted as part of Project 7 were 
distinct from interviews with emergency responders (NIST NCSTAR 1-8 2005), which were conducted in 
accordance with a separate agreement between the City of New York and NIST and between the 
PANYNJ and NIST. 



NIST NCSTAR 1-7, WTC Investigation 65 



Chapter 3 



This page intentionally left blank. 



66 NIST NCSTAR 1-7, WTC Investigation 



Chapter 4 

September 11, 2001, Before the Attacks 



4.1 



BUILDING POPULATION ON SEPTEMBER 11, 2001 



The total number of people inside WTC 1 and WTC 2 on September 1 1, 2001, is not known precisely, but 
it is necessary to make a good estimate of that number in order to provide context for understanding the 
evacuation of the buildings. On a typical Tuesday at 8:46:30 a.m., some businesses within the 
WTC complex would be largely staffed and operational, with others mostly empty, owing to a later 
corporate start time. In addition, September 11, 2001, was both the first day of the new school year for 
many children, as well as the date of primary election in New York City. Finally, tourists were not yet 
traveling to the observation deck in WTC 2 as it had not yet opened. These factors, among others, may 
have acted to limit the number of people who were at the WTC complex on that morning. 

The total number of building occupants is equal to the sum of survivors and decedents. Section 9.2 
contains an analysis of likely decedent locations at the time or aircraft impact. The response rate analysis 
for interviews presented in Section 3.2.2 leads directly to a projection of the number of people present in 
WTC 1 (8,900 ± 750) and WTC 2 (8,540 ± 920) on September 1 1, 2001, at the time of the first airplane 
impact. Table 4-1 indicates that the populations of WTC 1 and WTC 2 were similar (within statistical 
uncertainty intervals) and that 17,580 ± 1,180 individuals were inside the towers at 8:46:30 a.m. These 
numbers do not include any airplane passengers or crew, emergency responders, or bystanders. The total 
population was rounded to reflect uncertainty in the projection and decedent analysis. 



Table 4-1. Occupancy estimates on September 11, 2001, by 


tower. 


Estimate 


WTC 1 


WTC 2 


Total 


Estimated Total Population of 
Survivors 


7,470 


7,940 


15,410 


Estimated Number of Occupant 
Decedents 3 


1,462-1,533 


630 -701 


2,146-2,163 


Estimated Total Building 
Population 


8,960 


8,600 


17,560 



a. See Section 9.2 of this report for analysis of occupant decedent locations in WTC 1 and WTC 2. 

While Table 4-1 shows the estimated total population for WTC 1 and WTC 2, Table 4-2 shows the 
uncertainty calculations at different levels of statistical confidence. 36 The uncertainty is directly related to 
the number of interviews: more interviews completed results in less uncertainty in a projection (i.e., the 
number of occupants). Thus, the projection for WTC 1 has less uncertainty than the projection for 
WTC 2 (427 interviews and 376 interviews, respectively). 



' The standard error expressed in Table 4-2 assumes that the probability distribution is approximately normal (Gaussian). The 
standard error then defines the interval over which the actual population of each tower (as opposed to the estimated population) 
existed with an approximate level of confidence of 68 percent. The range of population expands as the uncertainty decreases, 
as shown by the 95 percent confidence (confidence limit at 5 percent) row in Table 4-2. 



NISTNCSTAR 1-7, WTC Investigation 



67 



Chapter 4 



Table 4-2. Occupancy estimate uncertainty. 




WTC1 


WTC2 


Total 


Number of Telephone Interviews 


427 


376 


803 


Standard Error (p) 


1.90% 


1.92% 


1.36% 


Standard Error (Total) 


750 


920 


1,180 


Confidence Limits at 5 % 


±1,470 


±1,790 


±2,320 



4.2 



OCCUPANT CHARACTERISTICS 



The survey data indicate that occupants of the WTC towers were twice as likely to be male as female 
(65 percent male (n=284) for WTC 1 and 69 percent (n=250) for WTC 2). As shown in Table 4-3 below, 
the average age of the occupants was mid-forties, with ages ranging from early twenties to mid-seventies, 
although one interviewee indicated that she attended the 90th birthday celebration for a fellow 
WTC evacuee (not included in the NIST sample) in 2003. The vast majority of respondents (92 percent 
(n=739)) spoke English as their primary language. It should be noted that some telephone contacts ended 
with a language barrier and that no interviews were conducted in any language other than English. 

Table 4-3. Age for telephone survey respondents in WTC 1 and WTC 2. 





WTC1 


WTC 2 


N b 


Valid 


439 


361 


Refuse 


1 


2 


Mean 


45 


45 


Median 


46 


44 


Minimum 


22 


21 


Maximum 


73 


74 



a. Mean and Median values are weighted. N, Minimum, and Maximum are unweighted. 

b. The sampling frame (badge list) identified 427 persons in WTC 1. 440 persons responded that they were in WTC 1 at 8:46 
a.m. on September 1 1, 2001 during the telephone interviews. 

Building occupants become familiar over time with a building, including the location of the emergency 
egress components and emergency procedures and protocols. The median residence time of the overall 
occupant population could thus be a predictor of how likely it was that an individual received training if 
the training was conducted every six months. 

Tenant and employee turnover at the WTC was not uncommon. Figure 4-1 shows the reported start dates 
for respondents in WTC 1 and WTC 2. In WTC 1, 4 percent (n=18) of the occupants had worked in the 
building since 1975. Further, 25 percent (n=l 10) had been working in the building prior to the 1993 
bombing, although only 16 percent (n=64) of the WTC 1 respondents were present on February 26, 1993. 
For WTC 1, 67 percent (n=287) of the occupants had started working in the building in the last four years 
(1998 - 2001). The mean residence time in WTC 1 was over 5.6 years, while the median was 2 years 
(half the respondents had been there two years or less, while half the respondents had been there longer 
than two years). 



68 



NIST NCSTAR 1-7, WTC Investigation 



September 11, 2001, Before the Attacks 



Occupant tenure in WTC 2 showed a similar trend. While only one respondent had worked in the 
building since 1975, 25 percent (n=91) of the respondents had been working in the building prior to the 
1993 bombing (with 16 percent (n=59) present on the day of the bombing). Another 51 percent (n=185) 
started working in the building at some point in the four years prior to the 2001 terrorist attack. The mean 
residence time in WTC 2 (n=360) was 5.9 years, while the median was 3 years. 




i i i i i i i i i i i i i i i i i i i i i i i i i i i 

YEAR STARTED AT WTC 1 



25.0%- 




20.0%- 




15.0%- 






10.0%- 




f] 


5.0%- 


„_„nnnnnn n nnnnnnnn W 





I I I I I I I I I I I I I I I I I I I I I I I I I I 
; is i 



YEAR STARTED AT WTC 2 



Note: Percentages are weighted. 

Figure 4-1. Employment start date at WTC. 

Overall, 7 percent (n=56) had a formal responsibility or special knowledge about the building. These 
respondents included fire safety staff, floor wardens, searchers, building maintenance, and security staff. 
Approximately 13 percent (n=105) of the respondents were employed by the PANYNJ, which may not 
necessarily imply a special knowledge of the building, as some PANYNJ employees had job duties 
related to functions outside the WTC. 

Some 6 percent (n=52) reported having a limitation that constrained their ability to evacuate. Both 
WTC 1 and WTC 2 had roughly the same fraction of the population who reported a mobility impairment 
(n=26 in each building). Six percent of the population of WTC 1 and WTC 2 corresponds to roughly 
1,000 people. The reported limitations included chronic illness, recent surgery or injury, obesity, elderly, 
heart condition, pregnancy, asthma, and other. Of these conditions, the most prevalent (n=20 of 52) was 
recent injury (in particular, severe knee and ankle injuries), followed closely (n=16 of 52) by reports of a 
chronic illness (such as cancer, leukemia, arthritis), or use of medications which hindered full mobility or 
cognitive ability. Four telephone interview respondents (of 52) reported being pregnant or having asthma, 
while three (of 52) reported having asthma- like conditions or indicated that age played a role in their 
ability to navigate the egress system. One person reported having a heart condition, while no telephone 
interviewee reported being blind, deaf, or requiring the use of a wheelchair. While the last three mobility 
impairments were not captured in the 803 telephone interviews, face-to-face interviews and published 
media accounts did provide information on people with these disabilities. A small number of respondents 
reported more than one mobility impairment. 



NISTNCSTAR 1-7, WTC Investigation 



69 



Chapter 4 



4.3 



PREVIOUS EXPERIENCE 



Of the WTC 1 occupants present on September 11, 2001, 16 percent (n=64) were also present during the 
1993 bombing. In WTC 1, 60 percent (n=38) of evacuees who evacuated both in February 1993 and in 
September 2001 reported that they evacuated immediately in 1993, while 30 percent (n=20) reported that 
they waited to evacuate in 1993, and 9 percent (n=6) did not recall. Most (95 percent (n=53)) who were 
able to recall their 1993 evacuation decision felt that they made the right decision, while 5 percent (n=3) 
did not believe they made the right decision. 

Similarly, 16 percent (n=59) of WTC 2 survivors on September 11, 2001, also evacuated in 1993. In 
WTC 2, however, only 75 percent (n=42) felt that they made the right decision in 1993 (compared to 
95 percent in WTC 1), possibly due to the fact that many more waited to evacuate in 1993 in WTC 2 
(69 percent (n=39)) than did so in WTC 1. Only 31 percent (n=17) of those who reported their decision 
evacuated immediately from WTC 2 in 1993. It should be noted that the bomb had a more significant 
impact upon WTC 1 in 1993. 



4.4 



OCCUPANT PREPAREDNESS 



Consistent with the NYC Building Code (Local Law 5 §C 19-1 62.2. a.4) and Occupational Safety and 
Health Administration (OSHA) regulations (29 CFR 1910.38, "Employee Emergency Plans and Fire 
Protection Plans"), 7 the tenants of the World Trade Center were required by the Port Authority to 
conduct regular fire drills. Further, they were required to appoint employee floor wardens and searchers. 
Overall, 66 percent (n=529) of WTC 1 and WTC 2 occupants reported participation in at least one fire 
drill in the twelve months immediately prior to September 11, 2001, and a significant proportion of 
occupants had taken part in two or more drills during that time. However, 17 percent (n=139) reported 
that they had not participated in any fire drills in the 12 months prior to September 11, 2001, and 
17 percent (n=135) did not know. Fire drill participation rates were similar between the two towers, as 
shown in Table 4-4. 

Table 4-4. WTC fire drills in 12 months prior to September 11, 2001. £ 



Number of Drills 


WTC1 


WTC 2 


None 


18%(n=78) 


17%(n=61) 


1 


13%(n=57) 


8 % (n=29) 


2 


21 % (n=90) 


24 % (n=88) 


3 


1 1 % (n=47) 


15%(n=53) 


4 


10%(n=44) 


9 % (n=32) 


5-11 


7%(n=31) 


9 % (n=32) 


1 2 or more 


3%(n=13) 


4%(n=13) 


Don't know 


18%(n=80) 


15%(n=55) 



a. Percentages are weighted, n values unweighted. 



While the NYC Building Code required fire drills every six months, OSHA regulations require fire drills at least annually. The 
PANYNJ required fire drills for all tenants every six months. 



70 



NIST NCSTAR 1-7, WTC Investigation 



September 11, 2001, Before the Attacks 



One of the goals of fire drill training was to make occupants aware of the location of the emergency exits. 
Of respondents who reported participation in a fire drill, 93 percent (n=490) indicated that they had been 
instructed about the location of the nearest stairwell. However, of the respondents who reported being 
shown a stairwell, 82 percent (n=432) did not enter or use the stairwell during the fire drill. Some 
17 percent (n=92) reported that they did use the stairs during a drill, while approximately 1 percent (n=5) 
reported not knowing whether they had used the stairs. Overall, more than half (51 percent (n=415)) of 
the occupants had never used a stairwell in WTC 1 or WTC 2 prior to September 11, while 48 percent 
(n=386) had used a stairwell. Two persons reported not knowing whether they had used the stairs 
previously. It should be noted, however, that Local Law 5, adopted in 1973, prohibits occupants from 
being required by building management or employers to enter or use a stairwell during a fire drill (1973). 
This prohibition may reflect social concerns regarding liability of required stairwell use by occupants and 
the economic costs of decreased employee productivity. In contrast, the City of Chicago, Illinois, 
(reacting in 2002 to the collapse of WTC) now requires twice-yearly fire drills, which "may conclude 
when all participating occupants have fully entered and have begun using designated stairwells." The 
City of Los Angeles, CA (Sixth Edition, 2002) in §57.33. 19. C, entitled "Emergency Planning and 
Evacuation Requirements for High-rise Buildings," requires that a "minimum of one fire drill annually on 
individual floors is mandatory. Total building evacuation is not required." 

Another goal of the fire drills was to introduce the floor warden system and evacuation procedures. Most 
occupants (82 percent (n=528)) with fire drill training were aware that there was a floor warden for their 
floor. Approximately 70 percent (n=557) of all occupants reported that they were aware of the evacuation 
procedures. When asked what those evacuation procedures comprised, however, answers varied 
significantly, including: wait in hallway for further instructions; do not use elevators, use stairs; meet at a 
designated site outside the building for a head count; or proceed down (varied number of) flights of stairs 
and wait. 

In general, most (66 percent (n=212) in WTC 1 and 60 percent (n=T67) in WTC 2) survivors who 
received fire drill training, reported that they found the training to be somewhat or very helpful to their 
evacuation experience on the morning of September 11, 2001. 



NISTNCSTAR 1-7, WTC Investigation 71 



Chapter 4 



This page intentionally left blank. 



72 NIST NCSTAR 1-7, WTC Investigation 



Chapter 5 

September 11, 2001, 8:46:30 a.m. 
Flight 11 Crashes into WTC 1 



"I heard the roar of the plane, looked out of the window [and] saw the 
plane halfway in the building. I jumped up and ran out into the hallway 
and screamed 'Everybody get out!'" Interview 1000749 (NIST 2004) 



5.1 



INTRODUCTION 



At 8:46:30 a.m. on Tuesday, September 11, 2001, American Airlines Flight 1 1, a hijacked Boeing 767, hit 
the north face of WTC 1, as shown in Figure 5-1. This impact resulted in a direct hit on seven floors, 
from 93 - 99, with additional damage extending several floors above and below the direct impact area. 
As Flight 1 1 crashed into the North Tower, all access to safety for those at or above the impact was 
destroyed, including both elevators and stairwells. The fate of over 1,300 occupants located above the 
91st floor of WTC 1 was sealed at that instant. This chapter focuses on occupant reaction during the 
initial moments after aircraft impact. 

As the aircraft struck the building, jet fuel on board ignited. Part of this fuel immediately burned off in 
large fireballs that erupted at the impact floors. Remaining fuel flowed across the floors and down 
elevator and utility shafts, igniting intense fires throughout portions of the buildings. A fireball killed or 
injured several occupants in the Concourse Level lobby. Despite the massive localized damage caused by 
the impact, as shown in the computer simulation in Figure 5-2 and Figure 5-3, the structure initially 
remained standing. 




Figure 5-1. WTC 1 impact, 8:46 a.m. 



NISTNCSTAR 1-7, WTC Investigation 



73 



Chapter 5 




Stairwell C 



njTUTfriJTrnjTfTLri 

U W U LU U 111 U W U 

TIES 

-V_J LEZ- 



Stairwell A 



Stairwell B 



[^r 



T7-T 



iiu 



LUULUU lSI 



lfunruriii 
juruuiuu 



i nn r*i n 



^1 



r 



WTC1 



© 



■ - '.' ' *-:<>.'j..--*V.'«.-iIy,<Li 




a 



n rnnfe'^i^p m runn 



lM\ 



WTO 



© 



Figure 5-2. Computer simulated impact damage to WTC 1 on floor 95 at 0.7 s after 

impact with stairwells superimposed. 



74 



NIST NCSTAR 1-7, WTC Investigation 



September 11, 2001, 8:46:30 a.m., Flight 11 Crashes Into WTC 1 




Floors 93 to 98 
Cumulative Damage 

Severe Floor Damage 

Flreproofing 

and partitions ' — ' 

Floor system . 1 

structural damage |_ 

Floor system 
removed '"• 



Column Damage 
Severed ^) 

Heavy Damage ^} 
Moderate Damage Q 
Light Damage 



r. ->■ 



» o.fi * 

4. A »Q A 



J. J. J.^ i. 



J. JJ. -U. XL XL XL 




Floor 93 



Floor 94 






- 


T TTT rj 


" r r t r t 


- 








iiiti 


n 




"\ 


" 


















- 












H- -'-'- 




* * M ' 


. 






t *- 








___ - ~ -. 














— * * ±&-i 








a & » 








- - * Pb 4 












" — 








„_ — — 








_- 




ii 








»- 










■ 


■"■ 










"" 


. 































Floor 96 



Floor 97 



Floor 98 



Figure 5-3. Calculated damage to floors 93 through 98 in WTC 1. 



5.2 



INITIAL OBSERVATIONS IN WTC 1 



The New York City 9-1-1 call center was quickly besieged with calls, primarily from citizens outside 
WTC 1, reporting that the World Trade Center was on fire. Some callers reported observing a 
commercial airliner intentionally hitting the building, and a few quickly concluded it was an act of 
terrorism. Other callers reported missiles and bombs, while some called 9-1-1 seeking to find out what 
was going on, possibly preventing or delaying people in danger from accessing the 9-1-1 system. For 
example, occupants from an upper floor of WTC 1, finding they were unable to connect to 9-1-1, called a 
business colleague outside the building, who in turn, was able to contact 9-1-1 and report their plight. 



3S 



The plane strike was immediately obvious to occupants of WTC 1 throughout the building. Even below 
the floors of direct airplane impact, building occupants knew that a significant event had occurred, and 
many witnessed significant fire, smoke, or building damage. The majority of survivors in WTC 1 felt the 
building move from the airplane impact (63 percent, n=277). Table 5-1 shows a summary of how the 
survivors became aware that something was wrong. The reported percentages were roughly constant 
throughout the building. For example, 60 percent (n=157) in the upper third of the building (floors 78 
and above) felt the building move, 62 percent (n=86) in the middle third (floors 44 to 77) felt the building 
move, while 64 percent (n=34) of the respondents in the lower third (up to floor 43) felt the building 
move. For some, this first cue was extreme; for others, less so. Still, it was clearly a significant event for 
all as the following quotes illustrate. 



New York City 9-1-1 Emergency Call Recordings, 2001. 



NISTNCSTAR 1-7, WTC Investigation 



75 



Chapter 5 



A survivor from a floor in the 70s felt tremendous movement in the 
building: "There were large vibrations, the building shuddering, the floor 
shaking violently. The initial explosion was a large cracking sound, and 
then boom. The building swayed heavily to the South the first time, and 
then a couple other times with decreasing severity." Interview 1000103 
(NIST 2004). 

Forty floors below the impact, the effect was immediate and significant 
as a survivor from a floor in the 50s notes: "There were creaking noises 
in the closet. I walked out into the hall and stood there. There was no one 
else in the hallway. I heard whooshing noises in the closet. The door 
blew open from the closet door, causing my chair to hit the desk" 
Interview 1000054 (NIST 2004). 

Thoughts of terrorism sprang quickly to the minds of many, particularly 
those who had survived the 1993 Bombing: "I felt the building sway. I 
knew it was really bad. My co-worker said, 'They did it again.'" Focus 
Group #4 (NIST 2004) 

Even occupants low in the building knew something major had occurred. 
A survivor from a floor in the 20s reported: "We felt the impact. The 
building swayed about seven times. Debris was falling down on the 
street. We gathered our belongings, I shut off the computer and headed 
towards the stairwell" Interview 1000559 (NIST 2004). 

Table 5-1. How survivors in WTC 1 became aware something was wrong on 

September 11, 2001. 



Cue 


Percent (n=440) 


Felt something (building moving, impact, shaking, swaying) 


63% 


Heard something (boom, crash, explosion, blast, roar, rumbling) 


30% 


Other, including saw a plane, smelled jet fuel, fell down/fell off chair, 
warned by someone 


7% 



Source: NIST WTC Telephone Survey Data. 

Even at ground level, awareness of the incident was immediate. One survivor who was in the Concourse 
of the building recounts this experience: 

"I was walking through the mall toward Tower One to get to the elevator. 
The lights flickered. I stopped in my tracks and looked around. I saw a 
brown cloud coming down the center corridor in the lobby, and I feared 
for my safety. The brown cloud had a heavy density and reached from 
floor to ceiling. It looked to me like it was both smoke and debris. It first 
came from the center corridor, but by the time it reached the revolving 
doors (a split second later) it seemed to come from every direction. At 
this point, the revolving doors exploded. They seemed to vaporize." 
Interview 1000046 (NIST 2004) 



76 NIST NCSTAR 1-7, WTC Investigation 



September 11, 2001, 8:46:30 a.m., Flight 11 Crashes Into WTC 1 



Like the Concourse Level, elevator lobbies throughout the building were particularly affected, 39 likely by 
excess jet fuel ignited by the crash pouring down the elevator shafts. 40 While only 3 percent (n=l 1) of the 
survivors reported seeing fireballs in their immediate area at the time of the airplane impact, the 
observations from the face-to-face interviews show the extreme nature of these events: 

A survivor from a floor in the 80s: "The entire corridor became an 
inferno outside our front door. Smoke began to enter our office. There 
was also debris falling. ... The fire on the corridor was at least 10 ft high, 
and it ran the . . . good length of the corridor. Then I saw a fireball come 
down the elevator shaft and blew the elevator doors. The fireball came 
right at me; it was a really bright color." Interview 1000055 (NIST 2004) 

A survivor from a floor in the 40s: "I saw the elevator in front me had 
flames coming out from it. The elevator was closed but the flames came 
from the front where the doors meet and on the sides. They reached 
about a foot and a half, with the flames standing from the floor to the 
ceiling. I saw a chandelier shaking; it was really moving. The corridor 
was dim. I also heard people screaming from the [nearby] floor. I felt the 
heat on my face and I thought that my eyebrows were going to get 
burned. Black smoke starting filling the corridor, it got really dense 
really fast." Interview 1000109 (NIST 2004) 

A survivor in the basement: "I saw a big bright orange color coming 
through the basement with the smoke ... A fire ball came shooting out of 
the basement door." Interview 100760 (NIST 2004) 

The elevator lobbies were not the only areas of the building damaged at the time of the airplane impact. 
Survivors noticed a range of damage and conditions throughout the building, from lost power to fire and 
smoke, to missing walls and floors. Table 5-2 shows observations at the time of first awareness. While 
some of the observations involve less severe phenomena (fallen ceiling tiles or flickering lights), others 
are more extreme, including collapsed walls, fire, and smoke. 

Observation of building damage during this initial awareness period were not as consistent over the height 
of the tower as the indications of the airplane impact. Table 5-2 presents a summary of observations 
reported during the telephone survey. While damage was more severe near the floors of impact, some 
damage was also evident at different locations lower in the building. Survivors provide a range of 
observations: 

A survivor from a floor in the 90s of WTC 1, just below the impact, 
recounts the severe damage on the floor: "In the hallway (from the 
bathroom to the elevator), there were no walls left (the wall board was 
blown off) and the bathroom seemed to be missing (the walls and the 
floor). There was a hole in the wall near the elevator (in the hall) and fire 
was coming up onto our floor through that hole." Interview 1000052 
(NIST 2004) 



39 



The majority of face-to-face interview respondents who observed fire inside the building, observed flames at or near the 

elevator shafts (NIST Interviews 2004). 

See NCSTAR 1-5 for further information about the consumption of jet fuel in WTC 1 and WTC 2. 

NIST NCSTAR 1-7, WTC Investigation 11 



Chapter 5 



A survivor from a floor in the 70s in WTC 1 : "To me everything seemed 
normal, all the ceilings were fine, the electricity was fine, and the air 
conditioning was also working." Interview 10001 18 (NIST 2004) 

A survivor from a floor in the 20s in WTC 1: "I was close to the 
windows. The windows were broken and I saw things from the office 
were going out the window." Interview 1000064 (NIST 2004) 

Table 5-2. Observations of building damage in WTC 1 when occupants first became 
aware something was wrong on September 11, 2001. 



Observation 


Percent (n=440) 


Fallen ceiling tiles 


17% 


Power outage/flickering lights 


17% 


Smoke 


10% 


Jet fuel 


8% 


Fire alarms 


8% 


Collapsed walls 


6% 


Other events, including fire, fireballs, 
injured people, fire sprinklers going off, 
extreme heat, debris 


45% 



Note: Total does not add up to 100 percent because respondents may have observed more than one event 

indicative of damage. 

Source: NIST WTC Telephone Survey Data. 

While a significant event, not all occupants felt their lives were in danger initially. Of the survivors in 
WTC 1, 41 percent felt their life was at risk, and 48 percent felt others lives were at risk, at first 
awareness. Only 4 percent of the survivors reported being injured by the attack initially, while only 
6 percent reported others being injured. 

Most of the survivors were with other occupants when the event occurred. One of every eleven survivors 
reported being alone at the time. Sixty-one percent were in a group of 10 persons or less, although two 
respondents reported being in a group as large as 400. The average reported group size was 23 persons, 
while the median group size was 7 persons. This suggests that a few reports of very large groups 
(33 respondents [8 percent] reported being a group of greater than 100 people) skewed the average. 

During the initial moments after WTC 1 was attacked, occupants above the 91st floor were trapped. A 
few occupants below floor 92 but near the impact region were alive, although trapped as well, some in 
their offices, others in elevators. Elevators were rendered inoperable. Occupants observed smoke, 
fireballs, damaged walls, fallen ceiling tiles, and injured colleagues on many floors throughout WTC 1. 



78 



NIST NCSTAR 1-7, WTC Investigation 



Chapter 6 

September 11, 2001, 8:47 a.m. - 9:02 a.m. 
Occupants React to the Attack on WTC 1 



"We gathered the group together to figure out what we should do. One 
exit was filled with smoke and it was dark; we tried to shut the door to 
keep out the smoke. The other exit was of no use, [so] we regrouped and 
went toward the main exit (towards the elevator). When the smoke was 
building up inside the office, I was more inclined to possibly break open 
the window and get some fresh air and wait for help; one or two of the 
other people insisted that we start evacuating [recognizing] that we were 
in very serious trouble. " Interview 1000137 (NIST 2004) 

Between 8:47 a.m. and 9:02 a.m., the time period after World Trade Center (WTC) 1 was attacked but 
before WTC 2 was attacked, one of the nation's largest building evacuations and emergency responses 
began to unfold. Occupants of WTC 1 , aware that something significant was happening in their building, 
were assessing their situation, performing necessary duties, and actively seeking a way out of the 
building. With no operational elevators available to the occupants, the three stairwells began to fill, not 
only with occupants exiting the building, but also with emergency responders entering the building. Near 
and above the floors of aircraft impact, the fire and smoke continued to spread, threatening the lives of the 
trapped occupants. Some occupants fell or chose to jump from the building. Building fire safety staff 
coordinated the response with incoming personnel from the Fire Department of New York, New York 
Police Department, Port Authority Police Department (PAPD), emergency managers from the City of 
New York, and Federal agents. At 8:47 a.m., an evacuation order for WTC 1 was broadcast over the 
vertical transportation channel (Z). The first radio communication regarding evacuation of WTC 1 and 
WTC 2 was issued by a senior PAPD officer at 8:59 a.m. Two minutes later, at 9:01 a.m., an instruction 
was issued to evacuate all WTC complex building (PANYNJ 2001a). NIST NCSTAR 1-8 contains a 
detailed chronology of all radio communications with respect to evacuation. 

Analysis of face-to-face interviews showed that many occupants of WTC 2, if afforded a view to the 
north or west, often chose to see for themselves what was happening to WTC 1. Once the significance of 
the event was verified, they were forced to decide, first, whether to evacuate or stay in place, and second, 
whether to evacuate using the stairs or the elevators. Building fire safety staff were deciding whether to 
order an evacuation of the occupants of WTC 2, taking into account the safety of WTC 2 occupants and 
what effect an evacuation of WTC 2 may have had on the evacuees from WTC 1 and on the incoming 
emergency responders (NIST 2004b). 

6.1 CONDITIONS WORSEN IN WTC 1 

Within minutes of the aircraft impact, occupants above the 91st floor began to assemble in groups of 
various sizes, often taking refuge in offices with access to windows. They also began to reach out for 
assistance, calling 9-1-1, family and friends, or colleagues. Electricity and phone service in the region 



NIST NCSTAR 1-7, WTC Investigation 79 



Chapter 6 



41 



directly impacted by the airplane was apparently disrupted. Electricity and phone service to floors 
above the impact floors was maintained, however, as evidenced by a number of 9-1-1 calls from the 100th 
to 106th floors during this time period. 



Fire and smoke raced upward. As early as 8:50 a.m., occupants on the 106th floors reported worsening 
smoke conditions for about 100 people on that floor, some of whom took refuge in a back office. 42 
Somewhat later, at 9:00 a.m., WTC Ch. 9 (PAPD Police Desk) received a call from a Windows on the 
World manager, who reported that floor 1 07 was "way too smoky" and most people had retreated to 
floor 106 (PANYNJ 2001a). 9-1-1 received reports of hazardous smoke conditions on 103, 104, 105, 
and 1 06 within ten minutes of aircraft impact. Each of those four floors reported having more than 
100 people trapped on the floor. For some, the conditions remained at least partially tenable during this 
time period. Others jumped or fell out of the building within minutes of the aircraft impact. Along with 
falling building debris, this created a hazardous situation for emergency responders, evacuees, and 
bystanders. 

Conditions for occupants below the 91st floor were deteriorating as well. More than one-third of the 
survivors reported seeing smoke after the initial airplane impact but before they left their initial floor to 
begin their evacuation, up from only 10 percent at first awareness. Eighteen percent of the survivors 
encountered the smell of jet fuel, up from 8 percent at first awareness. Table 6-1 shows a summary of 
these observations compared to those at first awareness. Observations of nearly all conditions increased as 
time progressed. 

Table 6-1. Observations of conditions in WTC 1 before beginning evacuation. 



Observation 


At Awareness 


During Interim Period 


Smoke 


10% 


35% 


Jet fuel 


8% 


18% 


Fallen ceiling tiles 


17% 


21% 


Power outage/flickering lights 


17% 


17% 


Fire alarms 


8% 


14% 


Collapsed walls 


6% 


10% 


Fire 


3% 


5% 


Other events 


45% 


48% 



Note: Total does not add up to 100 percent because respondents may have observed more than one event indicative of 

damage. 

Source: NIST WTC Telephone Survey Data. 



In WTC 1, a number of people below the impact zone were trapped on their floors, unable to either leave 
their offices or reach the stairwells. Prior to 9:02:59 a.m., trapped occupants requested assistance by 
calling 9-1-1 from an elevator, from scattered floors in the 10s, 20s, 30s, 40s, 60s, and from most floors in 



the 80s. Among those requesting assistance, heavy smoke, wall damage, and occupant injuries were 



41 New York City 911 Emergency Call Recordings, 2001. 

42 New York City 911 Emergency Call Recordings, 2001. 

43 91 1 Emergency Call Records, City of New York, 2001. 



80 



NIST NCSTAR 1-7, WTC Investigation 



September 11, 2001, 8:47 a.m. - 9:02 a.m., Occupants React to the Attack on WTC 1 



common reports. Figure 6-1 shows graphically the distribution of observed conditions in WTC 1 after 
initial awareness, but before beginning evacuation. The information to develop Figure 6-1 was compiled 
from every source available to the NIST investigation, including interviews, published accounts, 
transcripts of emergency communication channels, and emergency 9-1-1 calls. Note that the "?" denotes 
a floor where there was no information found to record the absence or presence of observations. Further, 
the absence of an observation on any floor does not positively exclude the presence of that condition as it 
may simply not have been reported. 





IIIIIIIIIII 


■ ■ 


I Tr 




■ ■ 


??????????? 


??????????? 


• 


• 


1 - 


— L 











■ 


IIIIIII III 



A 


' 




Wa 


#5 


<?i/ 


> 


fV 


% 


/-^ 


























1 


1 














































































































































































































































































































































































































1 


1 
































































































^ 
















* 


s? 


* 


»s 


*s 


« 











Figure 6-1. Observations of building damage after initial awareness but before 

beginning evacuation in WTC 1. 



6.1.1 



Activities and Information 



Occupants of WTC 1 engaged in a variety of activities prior to leaving their floor and beginning their 
evacuation, including exchange of information, gathering personal items, helping or searching for others, 
and fire fighting. Table 6-2 summarizes activities reported in the telephone survey. 

Many respondents used the time prior to beginning their evacuation to gather information about the event 
or to call family members. Others helped injured coworkers: 

A survivor from a floor in the 90s: "I heard a sound that sounded like a 
giant aluminum can being crushed and I felt the building tilt. I tried 
calling my company's home office but the line for long distance calls 
was not in service. I called home to test the phones and to let my family 
know that I was okay. I checked to see if our server was still up. I saw a 
man bleeding. I got a first aid kit and succeeded in halting the man's 
bleeding. We saw debris and smoke and decided it was time to get out. I 



NISTNCSTAR 1-7, WTC Investigation 



Chapter 6 



got my briefcase, a fire extinguisher, and four diet sodas, exited into the 
hallway and went towards stairwell C." Interview 1000052 (NIST 2004) 

A survivor from a floor in the 30s: "I used a telephone in the trading 
room to call my wife. I wanted to see if she had seen anything on the 
news and could tell me what was wrong. I called my wife within 
4 minutes of the impact of our building—and I got through okay. She 
wasn't there and I left a message. Because we saw the place of impact 
and fire coming out of the windows above us in our building, a 
co worker and I got our personal belongings and headed calmly to the 
stairwell." Interview 1000042 (NIST 2004) 

Table 6-2. Activities prior to evacuation reported in telephone survey 

by survivors from WTC 1. 



Activities Before Evacuation 


Percent Reporting the Activity (n=440) 


Talked to others 


70% 


Gathered personal items 


46% 


Helped others 


30% 


Searched for others 


23% 


Talked on telephone 


16% 


Moved between floors 


8% 


Fought fire or smoke 


6% 


Shut down computers 


6% 


Continued working 


3% 


Other activities 


25% 



Note: Total does not add up to 100 percent because respondents may have taken multiple actions. 
Source: NIST WTC Telephone Survey Data. 

Some occupants, on the other hand, started their evacuation almost immediately: 

An occupant from a floor in the 60s in WTC 1 : "It felt like the building 
was going to fall over. I grabbed my bag to leave the office floor. I was 
not waiting for anyone to tell me what to do." Interview 1000122 
(NIST 2004) 

An occupant from a floor in the 20s in WTC 1 : "I waited for building to 
stop shifting. I began to run straight out the nearest exit out of my office 
towards Stairway B. It was the nearest exit from my office and 
co-workers were just saying let's go this way." Interview 1000064 
(NIST 2004) 

While only 1 1 percent of the telephone survey respondents reported being given additional information 
about the event during this interim period without consciously seeking the information, 28 percent 
reported seeking such information. Table 6-3 shows the types of information received and sought by 
occupants. The majority of occupants received or sought information about the event, with a smaller 



82 



NIST NCSTAR 1-7, WTC Investigation 



September 11, 2001, 8:47 a.m. - 9:02 a.m., Occupants React to the Attack on WTC 1 



number looking for information on whether to evacuate or remain in the office. For those who sought 
additional information, 43 percent were unable to find the information they were seeking. 

Table 6-3. Information received and sought prior to beginning evacuation in WTC 1. 



Information Type 


Information Given (n=50) 


Information Sought (n=122) 


Information about what had happened 


57% 


81% 


Instructions to leave 


28% 


17% 


Instructions to stay 


17% 


12% 


Other, including information about what to 
do and to receive assistance in evacuation, 
don't know 


13% 


13% 



Note: Total does not add up to 100 percent because respondents may have observed more than one event indicative of damage. 
Source: NIST WTC telephone survey data. 

Occupants tried to obtain information through a variety of means -face-to-face conversation; telephone, 
television, or radio; e-mail or handheld devices; and from building announcements: 

A survivor from a floor in the 70s: "I walked to my desk and spoke on 
the phone to find out what happened. I went on the Internet and I was 
informed of what happened; also through telephone conversations. I 
thought it was necessary to look around. I walked around the floor with 
the fire warden; I also stopped, looked, and took some pictures. I was 
there in 93, and I wanted to wait for some directions from someone, 
through the speaker system, fire alarms, etc." Interview 1000576 
(NIST 2004) 

A survivor from a floor in the 60s: "During the time in which I was 
circulating in the northeast side of the floor, I heard a secretary say, 
'Aren't we supposed to wait for an announcement?' I saw other workers 
who were standing there talking and trying to assess what to do next. I 
went into the hall located between the stairwells A & C. People within 
the group helped each other make the correct decisions for evacuation." 
Interview 1000639 (NIST 2004) 

Twenty-seven percent of survivors felt they began their evacuation before the people around them. Not 
surprisingly, survivors' perceptions of risk of death to themselves and to others increased as conditions in 
the buildings worsened. Table 6-4 shows a comparison of survivors' perception of risk at the point of 
airplane impact and in the interim period before they left their floor to begin their evacuation. 

Table 6-4. Survivors perception of risk to self and others after airplane impact and 

prior to entering stairwell in WTC 1. 



Perception of Risk 


Others 


Self 


At awareness 


48% 


41% 


Interim 


63% 


52% 



Source: NIST WTC Telephone Survey Data. 



NISTNCSTAR 1-7, WTC Investigation 



83 



Chapter 6 



Although occupants felt at a heightened risk during this time, many occupants helped others before 
beginning their own evacuation; 20 percent of the survivors reported being helped by someone; while 
30 percent reported helping others. Table 6-5 summarizes the responses of the survivors who received 
assistance. 

Table 6-5. Sources of help used by occupants prior to beginning their evacuation 

in WTC 1. 



Source of Help 


Percent (n=87) 


Co-worker 


48% 


Police officer/firefighter 


16% 


Floor warden 


12% 


Manager/supervisor 


13% 


Other/don't know 


13% 


Stranger 


8% 



Note: Total does not add up to 100 percent because respondents may have received help from more than one person. 
Source: NIST WTC Telephone Survey Data. 

Requests for guidance or assistance were not uncommon. At 8:48 a.m., a mere two minutes after the 
plane hit WTC 1, an occupant from the 78th floor called the Port Authority Police desk requesting 
guidance: 

PAPD: Port Authority Police... 

Male: Yes, uh, we're on the 78th floor, at Hyundai Securities. Do we need to evacuate or 

not? 

PAPD: Right away. 

Male: Right now? 

PAPD: Right away. 

Male: Okay, which stairs do we take? 

PAPD: Uh, whichever is the easiest one nearest without too much smoke and everything. 

Try to get the best way down. 

Male: Well, there's. . .the hallways are full. . .full of smoke. 

PAPD: Okay. If you could find your way down one. . . 

Male: Okay, get out right now, right? 

PAPD: Right. Right, exactly. 

Male: Okay, bye. (PANYNJ 2001a) 

The telephone interviews revealed a variety of reasons which caused occupants decide to evacuate their 
floor. Table 6-6 summarizes the single predominant reason given by occupants for beginning evacuation 
in WTC 1. 

Some left due to observations of building damage or movement, others felt in danger, and still others left 
because co-workers left or told them to leave. Less than 1 percent said they left because they heard a fire 
alarm: 

A survivor from a floor in the 80s: "My boss told me that a plane came 
into the building. I was at a cubicle with no window view. I was 

84 NIST NCSTAR 1 - 7, WTC Investigation 



September 11, 2001, 8:47 a.m. - 9:02 a.m., Occupants React to the Attack on WTC 1 



screaming and crying, my boss came over to my location. We only had 
one door to enter or exit the office. The door was blocked with debris. 
We saw the ceiling caving in, but I don't recall any smoke or fire at this 
point. We began to dig out pieces of ceiling debris to open the door to 
exit." Interview 1000722 (NIST 2004) 

A survivor from a floor in the 60s: "The big boss, the treasurer, he stated 
"Get out now". I grabbed my pocketbook and started walking towards 
the stairwell." Interview 1000834 (NIST 2004) 

A survivor from a floor in the 20s: "I saw the floors in the hallway were 
twisted. The burning stuff outside the window was getting heavier and I 
decided that maybe I should look for a stairwell." Interview 1001667 
(NIST 2004) 

Table 6-6. Single reason given by survivors to begin their evacuation in WTC 1. 



Reason 


Percent (n=440) 


Building movement 


20% 


Afraid/felt in danger 


20% 


Was told to evacuate 


14% 


Friends/co-workers evacuated 


9% 


Saw debris 


6% 


Saw smoke 


5% 


Other, including saw fire, 1993 experience, saw/heard plane, people 
panicking, fire alarm going off 


26% 



Source: NIST WTC Telephone Survey Data. 

6.1.2 Emergency Response at the Fire Command Station, Lobby, WTC 1 

Within minutes of the initial impact, personnel from the Port Authority, building security, FDNY, NYPD, 
FBI, Secret Service, and representatives of Silverstein Properties (principal leaseholder) were assembling 
at the Fire Command Station in the lobby of WTC 1, as shown in Figure 6-2. The deputy fire safety 
director (a contractor from O'Conner Security) after assisting a woman injured by glass in the lobby, 
quickly began to receive and log calls from floor wardens on floors above the mid-rise area, including 
floors above the impact area. As the first calls came in, announcements were made to the affected floors, 
indicating, in general terms, 'We have received an alarm downstairs and the alarm is being investigated. 
Please stand by.' Any information from the floor wardens about the condition of the floor or injuries was 
passed to the Fire Department personnel nearby. As multiple floors were reporting incidents, the deputy 
FSD took down the floor numbers on a pad and paper and awaited the arrival of a supervisor. Within ten 
minutes, it was determined that the attack was a multiple-floor event. Therefore, consistent with 
emergency procedures, building-wide public address system announcements were made informing 
occupants to evacuate the building using the stairs and not the elevators. Initially, the evacuation script 



NIST NCSTAR 1-7, WTC Investigation 85 



Chapter 6 




Figure 6-2. Fire Command Station in WTC 1 on September 11, 2001. 



was used, but later the 
deputy FSD simply 
told people over the 
public address system 
not to use the 
elevators because they 
were crashing and to 
use the stairs to leave 
the building (Other 
Interview 03 
[NIST 2004b]). 
Unfortunately, the 
individual was 
unaware of the 
condition of the 22nd 
floor, where critical 
communications 



hardware in the hidden security command center lay in ruins, likely preventing any building-wide public 
address announcements from reaching the occupants (PANYNJ 2001a; NIST 2004b). After the fact, a 
person familiar with the operation of the building suggested that the fire alarm closet on floor 22 
destroyed the riser. NIST NCSTAR 1-4 has a more complete analysis of the fire alarm and public address 
system in WTC 1 and WTC 2. 

The damage on floor 22 was also reported by several emergency responders (NIST 2004) and was noted 
several times in the NIST analysis of the published accounts (Fahy and Proulx, 2003). NIST NCSTAR 1- 
8 (Chapter 4.4) contains additional information regarding the status of the 22nd floor command desk, as 
well as information flow between occupants, 9-1-1, and emergency responders. 



6.1.3 



Survivors Begin Their Evacuation 



Most, but not all building occupants began their evacuation of the WTC 1 before the WTC 2 was hit. 
Ninety-one percent of the survivors in WTC 1 reported beginning their evacuation before Flight 175 
struck WTC 2. At this point, nearly all observations of types of building damage had become more 
widespread than those at first awareness. Survivors recalled a variety of conditions on the floors as they 
left for the stairwells, ranging from significant damage to damage insufficient to deter the occupant from 
completing a routine task: 

From a floor in the 70s in WTC 1: "As I was leaving it didn't seem as 
bad as I thought in the office and I decided that I would just walk all the 
way down and reassess the situation and go back to the office if things 
were ok." Interview 1000129 (NIST 2004) 

From a floor in the 60s in WTC 1: "There was smoke and smell of jet 
fuel coming from the stairwell. I covered my nose/mouth with tissue. 
This smoke wasn't a lot; not to the extent that [you] could choke. The 
smoke was coming from the vents/corners of the stairwell." Interview 
1000036 (NIST 2004) 



86 



NIST NCSTAR 1-7, WTC Investigation 



September 11, 2001, 8:47 a.m. - 9:02 a.m., Occupants React to the Attack on WTC 1 



From a floor in the 60s in WTC 1 : "The first time I faxed my documents, 
they did not go through. I felt safe because I watched previous 
documentaries and I was informed that I was safe in my building. I went 
back to the fax machine." Interview 1000733 (NIST 2004) 

From a floor in the 50s in WTC 1: "I heard a chunk of ceiling fall and a 
woman screamed. We all stood and looked at each other and we tried to 
figure out what happened. We heard the cable snap in the freight elevator 
while we were talking, and the woman yelled "follow my voice". I 
followed the woman's voice to find where the stairwell was at to get out." 
Interview 1000054 (NIST 2004) 

From a floor in the 30s in WTC 1: "It [the stairwell] was the closest one 
to our office. I opened the doorway to the staircase... There was a lot of 
smoke and there was no one in it. I quickly closed the door." The 
occupant went to another stairwell down the hall to leave the floor. 
1000009 (NIST 2004) 

For consistency in evacuation measurement, time to begin evacuation was defined as the interval from 
first awareness to the time the respondent left his or her floor to begin evacuation. 44 On average, 
survivors in WTC 1 began their evacuation within 6 min. However, it is important to note that the 
statistical distribution of time to initiate evacuation was skewed in the direction of longer delays. In other 
words, while the most frequent response for survivors in all three zones in WTC 1 was one minute or less 
(referred to as the mode in statistics), and 50 percent of occupants had left their floor within 3-5 minutes 
(depending on zone), a few individuals took significantly longer (sometimes longer than 30 minutes) to 
start evacuating, thus disproportionately affecting the mean time to start evacuation. 

Table 6-7 summarizes the quartile, mode, and average times to start evacuation for survivors in WTC 1. 
Note that Table 6-7 separates the occupants into lower, middle, and upper floors based upon the location 
of the mechanical floors, which roughly divided the building into thirds. The reported times from the 
lower floors were not different than the reported times in the middle floors (tail probability from a log- 
transformed t-test comparing the two zones was 0.81). The upper floor evacuation initiation delay times, 
however, were statistically significant different when compared to both the middle and lower zones 
(significant at approximately the 99 percent confidence level). Thus, occupants nearer the impact area in 
WTC 1 delayed their evacuation for a longer period of time than occupants of the other two zones. This 
could have been due to the increased frequency of fire, smoke, building damage, and injured occupants on 
the upper floors, although that is only one explanation. A further discussion of evacuation initiation delay 
time and comparisons across regions of WTC 1 and WTC 2 is contained in Section 10.1. 



The time to begin evacuation was defined as the time while on the floor of origin due to the fact that, while many people 
decided to leave quickly, they often chose to perform several activities prior to actually entering the stairwell. Thus, time prior 
to entering a stairwell (or elevator) was a better measure of evacuation delay than a moment when the occupant 'decided' to 
evacuate, which may have been significantly prior to actually starting evacuation. 

NIST NCSTAR 1-7, WTC Investigation 87 



Chapter 6 



Table 6-7. Elapsed time (min) to initiate evacuation for survivors from WTC 1. 



Time for Survivors to 
Initiate Evacuation" 


25% 
Initiation 


50% 
Initiation 


75% 
Initiation 


Mode of 
Responses 


Average Time 
(min) 


Lower floors 
(Basement - 42) 


1 


3 


5 


1 


5.7 


Middle floors (43-76) 


1 


3 


5 


1 


4.8 


Upper floors (77-91) 


2 


5 


10 


1 


7.4 



a. Time to begin evacuation is the time interval from first awareness to the time the respondent left their floor to begin evacuation. 
Source: NIST WTC Telephone Survey Data. 

As shown in Table 6-8, survivors below the 92nd floor typically reported choosing the stairwell closest to 
them at the time. All three stairwells below the impact region were in use throughout the evacuation. 
Some found an appropriate stairwell quickly: 

From a floor in the 80s in WTC 1 : "The hallway was free of debris and 
well-lit. We (my boss, and co-workers, about ten to fifteen) went back to 
staircase C and proceeded to evacuate. The door on the staircase was not 
damaged. We entered staircase C. The staircase was well lit and fairly 
empty." Interview 1000108 (NIST 2004) 

From a floor in the 60s in WTC 1: "The building shook and I thought 
something tremendous had struck the building. I looked out the building 
to see what had happened. The fire alarm went off. I went to the stairwell 
(the exit) to evacuate the floor." Interview 1000025 (NIST 2004) 

From a floor in the 50s in WTC 1 : "It [the stairwell] was closest, and we 
had been trained in emergencies to only use the stairwells, never the 
elevators. We (four of us from my immediate office) exited down 
stairwell B." Interview 1000106 (NIST 2004) 

For others, finding an appropriate stairwell for evacuation was not always a straightforward process, as a 
survivor from a floor in the 30s recounts: 

"I opened the doorway to the staircase. There was a lot of smoke and 
there was no one in it. I quickly closed the door to not bring smoke into 
the floor. The group of people that was with me (about 1 people) started 
running back to the office. I began running after my coworkers and 
yelling at them to come back to find a different staircase. I was trying to 
do the right thing, and they were doing the wrong thing based on the fire 
drill training we had. The coworkers weren't listening so I let them go 
their own way and I went by myself back out to the hallway to find a 
different staircase. I walked down the north-south hallway back past the 
original stairwell (the one with the smoke in it) and made a right down 
the other hallway because I wanted to go with the crowd. There was a lot 
of traffic, so it took a little longer. After a couple of minutes, I went into 
the stairwell." Interview 1000009 (NIST 2004) 



NIST NCSTAR 1-7, WTC Investigation 



September 11, 2001, 8:47 a.m. - 9:02 a.m., Occupants React to the Attack on WTC 1 



Table 6-8 shows the percentage of occupants that chose each stairwell for evacuation, as well as their 
primary reason for selecting a stairwell. 

Table 6-8. Stairwell chosen for evacuation in WTC 1. 



Stairwell Used for Evacuation 


Percent 


Stairwell A 


17% 


Stairwell B 


25% 


Stairwell C 


19% 


Stairwell A or C 


10% 


Don't know 


17% 


Other, not applicable, used elevator 


12% 


Reason for Choosing Stairwell 




Closest one 


66% 


Followed others 


17% 


Was told to use 


12% 


Other exits blocked 


6% 


Other, including don't know, used before, best 
conditions, not applicable 


18% 



Note: Total does not add up to 100 percent because respondents may have given more than one reason for 

choosing their stairwell. 

Source: NIST WTC Telephone Survey Data. 



6.2 



INITIAL OBSERVATIONS AND REACTIONS FROM WTC 2 



Many of the occupants of WTC 2 quickly became aware that something significant had happened in 
WTC 1. Table 6-9 shows how survivors in WTC 2 became aware that something was wrong on 
September 11, 2001. Most occupants in WTC 2 heard, saw, or felt the event in WTC 1 (81 percent of the 
363 respondents). Others were made aware after a short time by coworkers, telephone, or the news media 
coverage. 

Table 6-9. How survivors in WTC 2 became aware that something was wrong 

on September 11, 2001. 



Observation 


Percent (n=363) 


Heard something (boom, crash, explosion, blast, roar, rumbling) 


51% 


Saw something (smoke or flames, plane, debris) 


19% 


Warned by someone around me 


13% 


Felt something (building moving, impact, shaking, swaying) 


11% 


Other, including contacted via phone, lights flickered, news media 


7% 



Source: NIST WTC Telephone Survey Data. 



NISTNCSTAR 1-7, WTC Investigation 



89 



Chapter 6 



For occupants near windows which faced north and west throughout WTC 2, what they saw made it 
instantly clear that the damage to WTC 1 was severe: 

A survivor from a floor in the 90s in WTC 2: "I heard a large noise that 
sounded like muffled dynamite and looked out the window which faces 
the East - Brooklyn Bridge or into Queens to see if the noise was 
connected to anything outside. I saw glittering paper which made no 
sense and thought it was part of a promotional event, as if some one was 
throwing confetti out of a plane. I headed in the direction of the noise 
and saw a gigantic red fire ball at the cubicle diagonally from my desk 
and smelled gasoline, which I later learned was jet fuel. I went back to 
my office to call my [spouse] to inform [him/her] that I was OK, and that 
I was leaving the building." Interview 1000001 (NIST 2004) 

A survivor from a floor in the 90s in WTC 2: "I felt a wave of heat; very 
high temperature. I went to the window on my floor to find out what was 
happening. I saw the flames/fire outside through the window (in Building 
One, just about the same floor as mine) and I saw everybody going to the 
manager. The manager instructed that, 'people without special 
responsibilities should evacuate the building.'" Interview 1000632 
(NIST 2004) 

A survivor from a floor in the teens of WTC 2: "I heard a loud roaring 
sound. I thought it was a window washer falling off its tracks. I walked 
into my office and looked out of the window. I saw people running in the 
plaza, away from building one. I saw paper and dust floating down from 
the sky. I saw the fireball coming from Building One. I thought it was an 
explosion of Windows on the World. However, I realized that there were 
still several floors above the explosion. Seeing the explosion and the 
panic in the plaza triggered my evacuation decision." Interview 1000922 
(NIST 2004) 

For some in WTC 2, however, the event was not as obvious. 

A survivor from the 100s in WTC 2: "A co-worker came to my office 
and said "There's a fire in the first building, we recommend that you 
leave." I grabbed my bag and packed up my belongings in order to leave 
the office. I grabbed my belongings for the reason in which I thought I 
was going to go to the gym & then would return to the office." Interview 
1000767 (NIST 2004) 

Although aware of the event, some occupants of WTC 2 were unsure of appropriate action to take since 
the event at this time was limited to WTC 1 . 

A survivor from a floor in the 50s in WTC 2: "I heard a loud, horrendous 
explosion. I turned and faced the source of the noise. I saw debris flying 
through the air outside the window. I saw large objects flying by, then 
reams of paper - some that were burning - like confetti. I looked up at 
One World Trade Center and saw a gaping hole and smoke. I called my 
wife and parents to reassure them, to tell them I was okay. Co-workers 
and I discussed what our course of action should be. We stood kind of 

90 NIST NCSTAR 1 - 7, WTC Investigation 



September 11, 2001, 8:47 a.m. - 9:02 a.m., Occupants React to the Attack on WTC 1 



paralyzed, undecided as to what we should do. The Bloomberg headline 
said a plane had hit WTC 1. We didn't know if we should close the desk 
and leave; if we should assume the day was over. We were looking for 
guidance." Interview 1000557 (NIST 2004) 

An occupant from WTC 2 at 8:49 a.m. also sought guidance from the Port Authority Police: 

PAPD: Port Authority Police... 

[Caller]: Yeah, this is [Caller], Securities Department from Morgan Stanley. 

PAPD: Uh-huh? 

[Caller]: Uh, what's the status right now as far as (overlap) 

PAPD: We're still checking. Everybody just get out of the building right now. 

[Caller]: All right. Have you guys announced an evacuation of Two? 

PAPD: We are trying to do that right now. 

[Caller]: All right, thank you. 

PAPD: All right? We are just advising everybody to get out of the building. 

[Caller]: All right, thank you, bye-bye. (PANYNJ 2001a) 

At 8:53 a.m., an occupant from WTC 2 called the Port Authority Police Department seeking advice. 

Male: Hi, um, I'm on the 95th floor of Two World Trade Center. 

PAPD: Yeah, just come on down anyway, sir. 

Male: Does that mean walk down the stairs? 

PAPD: You'd, be advised, right now at this time. 

Male: Should we evacuate all of our people? 

PAPD: Yes. Yes. Yes. Everybody. 

Male: Okay, thank you. (PANYNJ 2001a) 

6.2.1 Activities and Information - WTC 2 

Occupants of WTC 2 engaged in a variety of activities prior to leaving their floors and beginning their 
evacuation, including exchange of information, gathering personal items, helping or searching for others, 
and continuing work activity. Table 6-10 summarizes activities reported in the telephone survey. 

A survivor from a floor in the 60s of WTC 2 engaged in a variety of activities before beginning her 
evacuation. 

"I observed that the secretary was very upset. I went with [the secretary] 
and my manager to the lobby area to help care for [the secretary]. People 
were leaving the building and I didn't know when I'd be returning to the 
building, so I gathered my stuff and I went to the [rest] room. I saw 
flying things in the air and everyone was in a commotion to see what was 
going on, so I went back to the windows in the northern part of the 
building to find out what was going on. I heard the announcement . . . 
and I went back to my desk. [There] I made two phone calls to my 
mother and my [spouse] to reassure them that I was not affected and that 
I was safe. Two of my co-workers came into my office area and stated 
that they saw people jumping from Building 1 and that we should leave. 



NIST NCSTAR 1-7, WTC Investigation 91 



Chapter 6 



I gathered my stuff again and prepared to leave." Interview 1000877 
(NIST 2004) 

Table 6-10. Activities prior to evacuation reported in telephone 
surveys by survivors from WTC 2. 



Activities before Evacuation 


Percent Reporting the Activity (n=363) 


Talked to others 


75% 


Gathered personal items 


57% 


Helped others 


34% 


Searched for others 


32% 


Talked on telephone 


16% 


Moved between floors 


8% 


Shut down computers 


7% 


Continued working 


6% 


Fought fire or smoke 


1% 


Other activities 


20% 



Note: Total does not add up to 100 percent because respondents may have observed more than one 

event indicative of damage. 

Source: NIST WTC Telephone Survey Data. 

Another occupant from a floor in the 100s in WTC 2, who had only worked in the WTC for a couple of 
months, took relatively few actions before beginning evacuation. 

"I heard a female co-worker who had a window cubicle shouting 'Get 
out!' I turned around from my inner office, grabbed my purse, and 
walked out my office door. I ran to another co-worker and asked her 
what was going on. [The co-worker] didn't know and continued 
walking. I decided to follow the co-worker to [figure out] where to go 
and to find out what was going on. I followed [the co-worker] to a 
stairwell and began to go down the stairs." Interview 1000897 
(NIST 2004) 

While 2 1 percent of the WTC 2 telephone survey respondents reported being given additional information 
about the event during this interim period without actively seeking the information, 1 8 percent reported 
actively seeking such information. Table 6-1 1 shows the types of information received and sought by 
WTC 2 occupants. The majority of occupants received or sought information about the event, with 
smaller number looking for information on whether to evacuate or remain in the office. For those who 
sought additional information, 39 percent were unable to find the information they were seeking. 

Before WTC 2 was hit, information about the event affected occupants of WTC 2 in different ways. Even 
if an occupant heard that something happened to WTC 1 , he or she may have still felt safe in the building. 
For instance, an occupant beginning evacuation from a floor in the 50s in WTC 2 took to the stairs early, 
only to return to his/her desk. 



92 



NIST NCSTAR 1-7, WTC Investigation 



September 11, 2001, 8:47 a.m. - 9:02 a.m., Occupants React to the Attack on WTC 1 



"[After a few floors], I thought to myself, 'Why am I running?' I 
remembered that I was the fire warden for my floor." The occupant then 
took an elevator back to his/her original floor "to see if anyone else was 
there. I thought about the training and what I was supposed to do." 
After following the training procedure of calling security, the occupant 
"picked up the cell phone and called our [out-of-town] office to let them 
know that we were evacuating the building. I got ahold of them and they 
told me that a plane hit the building and to get out of there." Interview 
1001666 (NIST 2004) 

Table 6-11. Information received and sought prior to beginning evacuation in WTC 2. 



Information 


Information Given (n=77) 


Information Sought (n=64) 


Information about what had happened 


65% 


92% 


Instructions to leave 


26% 


17% 


Instructions to stay 


35% 


13% 


Other, including information to remain 
calm or a choice to evacuate or stay. 


10% 


5% 



Note: Total does not add up to 100 percent because respondents may have received or sought information from more than one 

source. 

Source: NIST WTC Telephone Survey Data. 

The occupant then proceeded to look around the floor for other occupants as well as make a phone call 
home to a family member. The occupant took an additional phone call when WTC 2 was hit, which 
prompted an immediate evacuation. Interview 1001666 (NIST 2004) 

However, another fire warden from a floor in the 40s of WTC 2 used observations from and media 
information about WTC 1 to begin evacuation immediately. 

After seeing paper flying outside the window and smelling gas, "I ran 
across the hall and came across the boss and immediately informed [the 
boss] that we should leave. I saw the TV and saw what had happened on 
CNN and was informed that a plane hit the building, Tower 1 . I went to 
the stairs." Interview 1000867 (NIST 2004) 

Twenty-four percent of survivors felt they began their evacuation before the people around them. Not 
surprisingly, survivors perception of risk to themselves and to others increased as conditions in the 
buildings worsened. Table 6-12 shows a comparison of survivors' perception of risk at the point of 
airplane impact and in the interim period before they left their floor to begin their evacuation. 

Table 6-12. Survivors' perception of risk upon airplane impact and 
prior to beginning their evacuation in WTC 2. 



Perception of Risk 


Others 


Self 


At awareness 


52% 


29% 


Interim 


67% 


42% 



Source: NIST WTC Telephone Survey Data. 



NISTNCSTAR 1-7, WTC Investigation 



93 



Chapter 6 



Although occupants felt a heightened risk during this time, many occupants helped others before 
beginning their own evacuation. Seventeen percent of the WTC 2 survivors reported being helped by 
someone and 34 percent reported helping others. Table 6-13 summarizes the responses of the WTC 2 
survivors. 

Table 6-13. How survivors received help prior to 
beginning their evacuation from WTC 2. 



Source of Help 


Percent (n=60) 


Co-worker 


56% 


Manager/supervisor 


15% 


Floor warden 


12% 


Other/don't know 


11% 


Police officer/firefighter 


7% 


Stranger 


5% 



Source: NIST WTC Telephone Survey Data. 

Even though the environmental cues were more obvious in WTC 1 than they were in WTC 2, 86 percent 
of the survivors in WTC 2 began their evacuation before their building was hit. Occupants gave a variety 
of reasons for beginning their evacuation, which are summarized in Table 6-14. 



Table 6-14. Single reason given by s 
evacuation from Wl 


urvivors to begin their 

rc2. 


Reason 


Percent (n=363) 


WTC 1 observations 


26% 


Was told to evacuate 


21 % 


Afraid/felt in danger 


17% 


Friends/co-workers evacuated 


11 % 


1 993 experience (thought was a bomb) 


6% 


WTC 2 building hit 


6% 


Other, including jet fuel, information seek, 
evacuating the building just felt like the right 
thing to do 


7% 



Source: NIST WTC Telephone Survey Data. 

Eighty-six percent to 91 percent of the occupants in WTC 2 began their evacuation before WTC 2 was 
hit. 45 Approximately one in four WTC 2 occupants began evacuation because of their observations of 
WTC 1 from their office windows, while a similar percentage (21 percent) were told to evacuate. 



45 86 percent was calculated by summing the number of people who reported starting their evacuation in less than 16 minutes, 
whereas 91 percent of people reported starting their evacuation prior to WTC 2 being attacked. These two questions were 
asked independently of one another during the telephone interviews. Note that while the discrepancy may be partially 



94 



NIST NCSTAR 1-7, WTC Investigation 



September 11, 2001, 8:47 a.m. - 9:02 a.m., Occupants React to the Attack on WTC 1 



An occupant from a floor in the 90s in WTC 2 left because of 
observations of WTC 1 : "[I] saw a gigantic red fire ball at the cubicle 
diagonal from my desk and smelled gasoline which I later learned was jet 
fuel. I went back to my office to call my [spouse] to tell [the spouse] that 
I was OK and that I was leaving the building." Interview 1000001 
(NIST 2004) 

Another occupant from a floor in the 30s in WTC 2 was told by company 
management that the occupant "should begin a self-evacuation, instead 
of waiting for the building management to tell us to evacuate. I went 
back to my desk to prepare to leave. I decided to follow the group that 
was heading for the stairwell." Interview 1000049 (NIST 2004) 

An occupant from a floor in the 90s in WTC 2 "heard a noise that 
prompted me to get up and look out the window. I saw people out of the 
corner of my eye, grabbing their bags and leaving. I turned and got my 
bag, with my pocketbook and things to leave my cubicle and follow the 
people [to the staircase]. I didn't have to investigate; I just left [because] 
I saw other people leaving." Interview 1000070 (NIST 2004) 



6.2.2 



Survivors Begin Their Evacuation 



On average, survivors in WTC 2 began their evacuation within 6 min. However, as with WTC 1 , the 
statistical distribution of time to initiate evacuation was skewed toward longer preparation periods. 
Overall, the most frequent response for survivors WTC 2 was one minute or less (although the middle 
floors had a mode of 5 min), and 50 percent of occupants had left their floor within 3-5 minutes 
(depending on zone), a few individuals took significantly longer (more than 30 minutes) to start 
evacuating, thus disproportionately affecting the average time to start evacuation. Table 6-15 
summarizes the quartile, mode, and average times to start evacuation for survivors in WTC 2. Note that 
Table 6-15 separates the occupants into lower, middle, and upper floors, based upon the location of the 
mechanical floors which roughly divided the building into thirds. The evacuation delay results may be 
biased for the upper floors in WTC 2, however, as only occupants who acted quickly to move below the 
78th floor before 9:02:59 a.m. could be interviewed. In other words, those who delayed for whatever 
reason, with few exceptions, did not survive. The impact of this potential bias was not quantified, but 
should be noted. A further discussion of evacuation initiation delay time and comparisons across regions 
of WTC 1 and WTC 2 is contained in Section 10.1. 

Table 6-15. Elapsed time (min) to initiate evacuation for survivors from WTC 2. 



Time for Survivors to 
Enter Stairwell 3 


25% 
Initiation 


50% 
Initiation 


75% 
Initiation 


Mode of 
Responses 


Average Time 
(min) 


Lower floors 
(Basement - 42) 


1 


4 


5 


1 


6.3 


Middle floors (43-76) 


2 


5 


10 


5 


7.1 


Upper floors (77-1 10) 


l b 


3 b 


5 b 


l b 


4.2 b 


a. Time to begin evacuation is the time interval from first awareness to the time the respondent left their floor to begin evacuation. 

b. The evacuation delay results may contain significant bias for the upper floors in WTC 2, as discussed in the text. 
Source: NIST WTC Telephone Survey Data. 



attributable to the uncertainty in time recollection by the occupants, the uncertainty intervals for the two percentages overlap at 
the 95 percent confidence level (plus or minus 5 percentage points). 



NISTNCSTAR 1-7, WTC Investigation 



95 



Chapter 6 



Occupants typically chose the stairwell closest to them at the time. Below the impact region, all three 
stairwells were in use throughout the evacuation. Table 6-16 shows the percentage of occupants each 
stairwell for evacuation and why that stairwell was chosen. 

Table 6-16. Stairwell chosen for evacuation in WTC 2. 



Stairwell Used for Evacuation 


Percent 


Stairwell A 


18% 


Stairwell B 


18% 


Stairwell C 


14% 


Stairwell A or C 


10% 


Don't know 


14% 


Other, including closest to the office, not applicable, 
used elevator 


27% 


Reason for Choosing Stairwell 




Closest one 


63% 


Followed others 


20% 


Was told to use 


10% 


Other exits blocked 


4% 


Other, including don't know, used before, best 
conditions, not applicable 


13% 



Note: Total does not add up to 100 percent because respondents may have given more than one reason for 

choosing their stairwell. 

Source: NIST WTC Telephone Survey Data. 

An occupant in the 80s of WTC 2 notes that colleagues "ran down the stairs in the South part of the office 
to evacuate the building. The staircase that I chose was the only one that I was aware of." Interview 
1000003 (NIST 2004) 

Another occupant in the 60s moved to the stairwell that was already being used. 

"I saw a person opening the door to a stairwell to the left of the reception 
area, near the men's room. We walked really fast to enter the stairwell to 
make our way down." Interview 1000526 (NIST 2004) 

An occupant in the 100s of WTC 2 was aware that he/she was assigned to use Stairwell B. 

"Due to the fire drill training, we were always told to use Stairwell B. I 
began going down Stairwell B in order to avoid elevators and to get out 
of the building safely." Interview 1000906 (NIST 2004) 



6.2.3 



Elevator Use in WTC 2 



Elevator usage by occupants played a significant role in reducing the total loss of life in WTC 2 on 
September 11, 2001. Sixteen percent of the occupants of WTC 2 used elevators for at least part of their 



96 



NIST NCSTAR 1-7, WTC Investigation 



September 11, 2001, 8:47 a.m. - 9:02 a.m., Occupants React to the Attack on WTC 1 



egress from the building; another 2 percent used elevators in the basement levels. Occupants higher in the 
building were somewhat more likely to use elevators than occupants lower in the building: subdividing 
the 16 percent who used elevators, 4 percent were from floors below the 44th floor skylobby, 5 percent 
from floors 46 to 73, and 7 percent from floors above the upper skylobby on the 78th floor. Expressed 
another way, 12 percent of survivors in the lower floors reported that they used an elevator, while 
15 percent of survivors from the middle floors reported using an elevator, and 33 percent of survivors 
from the upper floors reported that they used an elevator to evacuate WTC 2. Interpretation of the latter 
percentage may be biased, however, as occupants who used stairwells or delayed evacuation may have 
been disproportionately trapped above the impact region after 9:02:59 a.m. when compared to occupants 
who initially chose to evacuate using elevators over stairwells. 

For some occupants, the physical challenge of navigating the stairwells was daunting enough to 
precipitate rapid use of the elevator system. 

An occupant from the 90s: "I think perhaps a survival instinct prompted 
my decision. I grabbed my briefcase and walked fast toward the 
elevator. I saw someone who needed to use crutches just as the elevator 
door was closing on the floor. We opened the door for him manually; 
the 'open door' button never functioned properly. I was taking a new 
medication and knew I should not walk down the stairs, so I took the 
elevator to the 78th floor and then took another elevator to the lobby." 
Interview 1000553 (NIST 2004) 

For others, physical challenges presented themselves while in the stairwells. Even before WTC 2 was 
attacked, Port Authority communications (Channel 27) indicate that an occupant required assistance in the 
stairwells at 9:00 a.m. (PANYNJ 2001a) Others were able to switch egress modes from the stairwells to 
the elevators, although not without some difficulty, as one occupant from the 40s demonstrated: 

"My leg was hurting (it was a pre-existing condition) and I had 
bronchitis. The security guard wouldn't let me go down the express 
elevator, [so] this man agreed to come to my aid. [We] started our 
descent on the steps and continued to the 40th floor where he knew of 
another elevator bank. The man who was helping me announced to the 
people around us that he was helping somebody (me) and that we would 
stop at each landing and get out of their way so that they could pass us 
easily. We exited the staircase on the 40th floor and found and entered an 
elevator. Two women joined us because they thought we knew what we 
were doing. We rode the elevator from the 40th floor to the lobby. It 
made two stops before expressing to the lobby." Interview 1000048 
(NIST 2004) 

Some occupants were successful in using an elevator to get to the lobby, but were turned around by 
building security, presumably in an effort to minimize the impact WTC 2 occupants would have posed on 
the evacuation of WTC 1 occupants: 

"I was in an elevator going up to my floor when Building One was hit. 
When I got to my floor, we took an elevator back down to lobby, but the 
guard sent the elevator full of people back up to the floor." Focus 
Group #3 (NIST 2004) 



NIST NCSTAR 1-7, WTC Investigation 97 



Chapter 6 



6.2.4 Announcements in WTC 2 

At 8:58 a.m., WTC Channel 15 (NYC EMS Direct Line) recorded a statement from an unknown source, 
stating, "I want to start a building evacuation, Building One and Building Two." (PANYNJ 2001a). The 
timing was proximate with announcements in WTC 2, made from the lobby fire command station, first 
contradicting the Channel 1 5 statement, then supporting it. 

At 9:00 a.m., approximately three minutes before the second attack, the first building-wide public address 
system announcement was made to the occupants of WTC 2. Synthesizing the content of the 
announcement as reported by many interview respondents indicates that the general announcement went 
as follows: 

There is a fire condition in WTC 1. WTC 2 is secure. Please return to 
your offices. 

While some occupants interpreted the announcement as less of an instruction and more of a suggestion 
that it was safe to return to their offices, others interpreted the announcement as an instruction to remain 
where they were or to return to their offices. No respondent reported that the announcement told them to 
evacuate. 

Approximately two minutes later, at 9:02 a.m., one minute before WTC 2 was attacked, a follow-up 
announcement was made, contradicting the previous announcement. Recorded in the background of an 
answering machine recording an occupant calling home from a floor in the 90s and subsequently provided 
to the interviewer during the interview, the announcement indicated that it was now permissible to initiate 
an evacuation: 

"May I have your attention please. The situation is in Building 1. 
However, if conditions on your floor warrant, you may wish to start an 
orderly evacuation." Interview 3000001 (NIST 2004) 

The 9:02 a.m. announcement was also noted by an occupant catalogued in the collection of published 
accounts (Fahy and Proulx 2003). A survivor in the 70s of WTC 2 was still on his or her floor when the 
9:02 a.m. announcement was given over the public address system. 

"I stayed on the floor [to] wait until everyone was cleared out. Most of 
the people on my floor evacuated using the elevators, which they took to 
the skylobby. It seemed to me [that] I would be in more danger by going 
out. I heard the phone ringing off the hook and my boss and I were 
going from phone to phone answering them to reassure the families that 
everyone was ok." Interview 1000524 (NIST 2004). 

It was then that United Airlines Flight 175 crashed into WTC 2. 



98 NIST NCSTAR 1-7, WTC Investigation 



Chapter 7 

September 11, 2001, 9:02:59 a.m. 
Flight 175 Crashes into WTC 2 



"On the phone, [the occupant from a floor in the 90s in WTC 2] kept 
describing the people jumping out of WTC 1. I could not hear any noise 
in the background; it was quiet with some slight conversations going on. 
When the second plane hit WTC 2, I then heard people screaming, "No, 
no. Oh my God, no!" Screaming and then silence. People's sounds 
disappeared. I could hear 'whopping noises. ' The sounds were similar 
to wind blowing and I could hear the fire alarms going off. Then, 
nothing else. Just silence. " Interview 300003 (NIST 2004) 



, im unni i irtHii i in'" ^^ * 



Stairwell A 



E 

E3E3E3'^U 
2=3 = £J=> ■ ■•• 



.A 



n2Tl 






Stairwell 
B * 



At 9:02:59 a.m., a second hijacked Boeing 767-200ER, United Airlines Flight 175, struck WTC 2, 
damaging nine floors, from 78 to 84, as shown in Figure 7-1. By the time WTC 2 was hit, 21 percent of 

the eventual survivors had already exited WTC 1, 
and 4 1 percent of the survivors had already exited 
WTC 2. With the second attack, occupants in 
WTC 2, like those from WTC 1, began a full- 
scale building evacuation. 

Alarm tones and public address announcements 
were overheard in 9-1-1 audio records from 
various floors from floors 82 to 1 05 after the 
impact, indicating that some power was available 
to most of the building. 

As when WTC 1 was struck, a significant surge 
in 9-1-1 calls occurred when WTC 2 was struck, 
as bystanders and people watching television 
called to notify authorities that they witnessed the 
event. Many witnesses immediately reported that 
a deliberate terrorist attack was underway. Once 
again, the WTC 2 attack was described to 
operators in a variety of ways, from a large 
military aircraft to a bomb, to a commercial 
aircraft. 46 




Stairwell C 



^ 




© 

WTC 2 



Figure 7-1. WTC 2 damage (computer 

simulated) at 0.62 s after impact with 

stairwells superimposed. 



New York City 911 Emergency Call Recordings, 2001. 



NISTNCSTAR 1-7, WTC Investigation 



99 



Chapter 7 



In WTC 1 , many occupants only had passing knowledge of the second plane strike from cell phone or 
electronic communications received by other occupants within the stairwells during their egress: 

A survivor from the 50s in WTC 1: "Very soon after entering the 
stairwell, someone with a Blackberry communication device received 
word that a corporate jet had hit the building. Later the info was clarified 
as to what kind of plane it was. Later a report of the second plane hitting. 
We could feel the building shake a little, not that strong though." 
Interview 1000015 (NIST 2004) 

Other occupants sensed that something happened, but had no idea what the source of the disturbance was. 

A survivor in the 60s in WTC 1 : "I felt the air clear up due to a big gush 
of wind that cleared the smoke (this happened when Building 2 was hit 
by the plane)" Interview 1 000 1 00 (NIST 2004) 

Occupants of WTC 1 with a view of the plaza observed large pieces of debris, some flaming, crashing 
down onto the central plaza. Evacuees from Stairwells A and C used escalators on the east side of the 
Mezzanine to proceed down to the lobby. The escalator access point, which often had a group of people 
waiting their turn to use the escalators, had a direct view of the debris, as shown in Figure 7-2. 




Figure 7-2. East-looking view through WTC 1 lobby as debris from WTC 2 impact travels 

past lobby windows onto plaza. 



100 



NIST NCSTAR 1-7, WTC Investigation 



September 11, 2001, 9:02:59 a.m., Flight 175 Crashes into WTC 2 



As occupants exited the building onto the street or plaza level, many were unknowingly walking into 
danger, as debris rained down from the impact region in WTC 2. An occupant ofWTC 2 from the 100s, 
having used an elevator to get to the lobby, exited WTC 2 at 9:03 a.m.: 

"I was guided by what I think were police officers who directed the 
crowd. I imagine they were guiding [us] away from where the debris 
was falling off Building 1. I walked out straight ahead towards the 
Liberty St. exit. Outside of Building 2, debris was falling down when the 
second plane hit. A piece of metal went into my arm. I went into shock 
due to severe blood loss [and] was assisted by NYPD, which [took] me 
to the hospital." Interview 1000563 (NIST 2004) 

Due to the angle of the aircraft impact, large regions of survivability remained on some of the floors 
directly damaged by Flight 175. Figure 7-3 shows the individual and aggregated damage areas for floors 
78 - 83 in WTC 2. See NIST NCSTAR 1-2 and NIST NCSTAR 1-6 for additional building damage 
information. From the time of impact and for the next 20 minutes, occupants called 9-1-1 from the 
impact region and above to request assistance. 47 




Floors 78-83 
Cumulative Damage 

Severe Floor Damage 

Flreprooflng 

and partitions ' — ' 

Floor system i 1 

structural damage | | 



Floor system 
removed 



Column Damage 
Severed Q 

Heavy Damage ^j 
Moderate Damage Q 
Light Damage 





j~7 7 7 7 7 ~ 7 " 7 " " 


"T i T T.I 


















— 







l'i i x 


n — 


" 






















- 


L 

A.±. 














J— 


, — _ 


4 -I 1 — 1 — — 


_ 










„ 




J-i-4 


_ 


_ 






>_. — — - 


■ 




n 


.* *• 




■ 






ff 1 ' 


- 


' 










- 
















in 















































B"t 


"j!l 


4 





' 


















- i j. * 






































» * 






















- 




j j- 








■r 


■ 




■J 






~ 


































> P° 0( 










■ < 


1 




_ 


m*. 














- 












" 






■ *" 










.... 

















Floor 78 



Floor 79 



Floor 81 



Floor 82 





























■ 










- 







il'i M i!iJ 


TTS 1 






































































- 








- 








Xi , 




- 
















1 ~ x ~wQ 








I 


-.','.'.', 


■yc= 




- 




>r i'tt' 


, ;:; Ei 




■ 


-• 





Floor 80 




ma 



A H * w - 



IUMB 



Floor 83 



Figure 7-3. Calculated damage to floors 78 through 83 in WTC 2 (see NIST NCSTAR 1-6). 



New York City 911 Emergency Call Recordings, 2001. 



NIST NCSTAR 1-7, WTC Investigation 



101 



Chapter 7 

Standing on a floor in the 70s in WTC 2, a survivor described the immediate aftermath of the aircraft 
impact: 

"I heard this deep thud and the ceilings and walls started to crumble. I 
grabbed my laptop, ran out of my office. I ran towards Stairwell A 
[because] this was the closest stairwell to my office. One of my 
colleagues opened the door to Stairwell A, but we didn't go in and turned 
around and went south. Stairwell A (the NW stairwell) was pitch black, 
filled with heavy black smoke. There was falling debris from the ceiling 
making the other two stairwells inaccessible [so we] ran back to 
Stairwell A [because] that was only stairwell we had access to. There 
was a lot of soot and dust in the air on the way back to Stairwell A. I had 
to cover my mouth and nose with my shirt to be able to breathe." 
Interview 1000625 (NIST 2004) 

A survivor from a floor in the 100s in WTC 2 made it to the 78th floor skylobby just prior to the impact 
of flight 175: 

I was walking down the stairs, and got off at 78 (stairwell C). I 
encountered a lobby full of people (found some people from my group). 
The plane hit the building. I went flying and I landed on my right arm. 
When that happened my body turned over onto its left side and was 
sliding towards the elevators that were being repaired. I thought I would 
go down into the elevator shaft (since the door was slightly ajar and you 
could see fire coming up and I thought that it would be the end). I 
stopped and I called out to my friends and they called back with their 
location. I walked over to them climbing over dead bodies. We made an 
assessment of what had happened (my boss was dead, other man had 
broken legs and my coworker was missing). I got up and was walking in 
the direction of the plane. As I was walking there I came across people 
who were upset and I was climbing over bodies and I came to realize 
there was no communication desk." Interview 1000562 (NIST 2004) 

The majority of WTC 2 occupants were not on their usual work floors when WTC 2 was attacked; most 
people had initiated or completed evacuation. Therefore, relatively few observations of damage to the 
floors in WTC 2 were reported. Figure 7-4 shows a summary of the reported damage to floors in WTC 2. 
The information to develop Figure 1-^ was compiled from every source available to the NIST 
investigation, including interviews, published accounts, transcripts of emergency communication 
channels, and emergency 9-1-1 calls. Note that the "?" denotes a floor where there was no information 
found to record the absence or presence of observations. Further, the absence of an observation on any 
floor does not positively exclude the presence of that condition as it may simply not have been reported. 



1 02 NIST NCSTAR 1-7, WTC Investigation 



September 11, 2001, 9:02:59 a.m., Flight 175 Crashes into WTC 2 







4. 








!? 


? 


■> 


■> 


? 


■> 


? 


? 


3 


? 


7 


-> 




X 


X 


X 


g 


X 


g 


§ 


X 


g 


X 


X 




X 


X 


X 


S! 


X 


s 


g 


X 


§ 


X 


X 




























? 


? 


? 


? 


? 


? 


? 


? 


? 


? 


? 










































































? 


-> 


? 


? 


9 


? 


? 


? 


? 


? 


? 






■> 


■5 


? 


-> 


? 


? 


? 


? 


■? 


-> 


















































o 


■p 


? 


■> 


? 


? 


•> 


? 


? 


-> 


























7 


? 


? 


•> 


? 


? 


? 


? 


? 


f 






-> 


■> 


? 


•> 


? 


? 


? 


? 


? 


-> 






-> 


■> 


? 


-> 


? 


? 


■> 


? 


7 


-> 






























? 


7 


? 


•> 


? 


? 


7 


? 


7 


? 






? 


? 


? 


•> 


? 


? 


? 


? 


■? 


? 






? 


? 


? 


? 


? 


? 


? 


? 


? 


? 






-> 


■p 


? 


■> 


? 


? 


■p 


? 


■p 


9 


























o 


? 


? 


■> 


? 


? 


? 


? 


? 


■? 


























? 


? 


? 


? 


? 


? 


? 


? 


? 


? 


















































































































? 


? 


? 


? 


? 


? 


? 7 1 7 


? 


? 



























































































/# 


' 


/v 


/■::■ 


■ : y 


Pv 






















































- 


= 


- 














































































' 


7 


■> 


~r 


■> 


7 


> 


■> 


? 


' 


' 


» 


> 


■> 


■> 


' 


7 


1 


» 


' 


' 


7 


■> 


> 


■> 


■> 


5 


' 


' 


? 


? 


' 


' 
























? 


' 


? 


■> 


' 


' 


■> 


» 


» 


' 


' 


3 


£ 


$ 


x" 


£ 


g 


Si 


S 


X" 


3 


g 




LJ. 



Figure 7-4. Observations of damage from tenant spaces in WTC 2. 



NISTNCSTAR 1-7, WTC Investigation 



103 



Chapter 7 



This page intentionally left blank. 



104 NIST NCSTAR 1-7, WTC Investigation 



Chapter 8 

September 11, 2001, 9:02:59 a.m. - 9:58:59 a.m. 
56 Critical Minutes 



With the attack on World Trade Center (WTC) 2, the events in both WTC 1 and WTC 2 unfolded 
similarly, with survivors in both buildings trying to or continuing to evacuate the buildings. For those 
trapped above the plane strike in WTC 1, there was no way out. For those above the impact in WTC 2, a 
single intact stairwell provided a critical lifeline for at least 18 occupants fortunate enough to both find 
and use it. 48 

In both buildings, stairwells became not only a means of egress for occupants, but also the only 
significant route into the building for emergency responders. Interview analysis found that, prior to the 
collapse of WTC 2, the six stairwells in the two towers were relatively passable below the impact regions. 
While many occupants indicated that the conditions in the stairwells were better than the conditions 
experienced during the 1993 bombing, water, crowding, and the smell of jet fuel were cited as stairwell 
obstacles. All the while, building officials and emergency responders tried to cope with an incident of 
enormous scope and provide for as orderly an evacuation as possible. 

By the time WTC 2 was hit, 91 percent of the survivors in WTC 1 and 87 percent in WTC 2 had begun 
their evacuation. Over 20 percent of the survivors in WTC 1 and more than 40 percent in WTC 2 had 
already left the buildings at the time of the second airplane strike. Table 8-1 summarizes when survivors 
began their evacuation and when they left their respective building. 

Table 8-1. Estimated times for survivors leaving the building in WTC 1 and WTC 2. 





WTC 1 (n=440) 


WTC 2 (n=363) 


Begin 
Evacuation 


Leave Building 


Begin 
Evacuation 


Leave Building 


Before WTC 2 hit 


91 % 


21 % 


87% 


41 % 


After WTC 2 hit, but before 
WTC 2 collapse 


7% 


67% 


13% 


58% 


After WTC 2 collapse 


0% 


9% 


n.a. 


n.a. 



Key: n.a., not applicable; question not included in telephone survey. 

Note: Numbers may not total to 100 percent. Some survivors were unsure when they began their evacuation and/or when 

they left the building. 

Source: NIST WTC Telephone Survey Data. 



See Section 8.1 for further information. 



NISTNCSTAR 1-7, WTC Investigation 



105 



Chapter 8 



8.1 



SURVIVORS FROM AT OR ABOVE FLOOR 78 IN WTC 2, AFTER IMPACT 



Table 8-2. Location of WTC 2 survivors at or 
above floors of impact at 9:03 a.m. 



Above the floors directly damaged by 

Flight 175 in WTC 2, smoke traveled rapidly 

upwards. Within twelve minutes of WTC 2 

impact, a smoke layer was descending on 

floor 105, 49 where multiple callers reported 

that between twelve and one hundred people 

were trapped. 50 Through face-to-face 

interviews and analysis of published accounts, 

however, NIST identified 18 individuals at or 

above the floors of impact in WTC 2 after 

9:02:59 a.m. who escaped. The majority of 

these survivors were on the 78th floor, often 

referred to as the skylobby, waiting for an 

express elevator to the lobby or choosing a 

course of action just prior to the impact. 

Table 8-2 shows the location of the 18 

individuals at 8:46:30 a.m., as well as 

9:02:59 a.m. For the hundreds of occupants 

who remained above the skylobby when 

WTC 2 was hit, two of the buildings three 

stairwells had been destroyed by the impact. 

Only the staircase furthest away from the 

plane's impact was passable (Stairwell A) and 

even this stairwell was severely damaged in 

places. According to eyewitness accounts, 

there was drywall and other debris scattered 

over the stairs, water was running down the 

stairs, presumably from the sprinkler system 

(Murphy 2002), and smoke filled the stairwell 

(Adler 2002). In the region near the airplane 

impact, the stairwell walls had collapsed, and 

occupants had to crawl over or under the 

debris in order to descend below the 78th 

floor (Adler 2002). According to NIST 

interviews and several published accounts, the 

conditions in the stairwells then improved significantly below the 78th floor. 

Published accounts indicate at least two individuals were below the 78th floor in WTC 2 at the time of 

impact, but ascended in order to help injured or trapped colleagues (Fahy and Proulx 2003; 

Murphy 2002). Both individuals were building occupants, and one individual was reported to have had 





Location at 8:46 a.m. 
(WTC 1 Impact) 


Location at 9:03 a.m. 
(WTC 2 Impact) 


1 


Floor 103 


Floor 78 


2 


Floor 103 


Floor 78 


3 


Floor 102 


Floor 78 


4 


Floor 101 


Floor 78 


5 


Floor 100 


Floor 82 (Stairs) 


6 


Floor 100 


Floor 78 (Elevator) 


7 


Floor 100 


Floor 78 


8 


Floor 97 


Floor 78 


9 


Floor 97 


Floor 78 a 


10 


Floor 86 


Floor 78 


11 


Floor 86 


Floor 78 


12 


Floor 84 


Floor 84 


13 


Floor 84 


Floor 84 b 


14 


Floor 84 


Floor 84 (Elevator) 


15 


Floor 81 


Floor 81 


16 


Floor 81 


Floor 81 


17 


Floor 79 


Floor 78 


18 


Floor 78 


Floor 78 



While this individual did evacuate the building, the occupant died 
several days later as a result of injuries sustained on September 11, 
2001. 

After impact, this occupant ascended to floor 91 for a period of time 
before finally making it below the impact zone. 



New York City 911 Emergency Call Recordings, 2001. 
' New York City 911 Emergency Call Recordings, 200 1 . 



106 



NIST NCSTAR 1-7, WTC Investigation 



September 11, 2001, 9:02:59 a.m. - 9:58:59 a.m., 56 Critical Minutes 



an emergency response background. Neither individual successfully evacuated, although several 
survivors credited one of the individuals with saving their lives. 

8.2 PROGRESS OF EVACUATION BELOW THE IMPACT REGION 

For nearly all occupants below the floors of impact, the stairwells became the only means of egress from 
both towers. For some the egress went smoothly; for others it was far more difficult. Deteriorating 
conditions, coping with crowded stairwells, and the sheer physical effort required to descend dozens of 
flights of stairs presented a challenge for many occupants. 

A survivor who was on the 78th floor skylobby in WTC 2, then used the 
stairs down from there: "People were having general conversations, 
seemed calm, and walked at a steady pace, no sense of panic." Interview 
1000825 (NIST 2004) 

A survivor who began on a floor in the 30s of WTC 1: "On about the 
29th or 30th floor, as I was leaning against the wall letting an injured 
person pass, I felt the wall heave (which apparently was due to the 
second plane hitting the other tower). This was about 9:10. On the stairs, 
near the 28th floor, there was a pile of shoes that accumulated from 
people kicking them off. Some of the people around me were tripping on 
them and warning others to watch out for them. Our speed of descent 
was very slow. There were many firemen clogging the stairs and spilling 
out onto the 28th floor. They were resting and taking care of one who 
was sick—was having a heart attack. They were also carrying much 
equipment. All of this was happening at about 9:20 am." Interview 
1000042 (NIST 2004) 

Occupants also sought information from emergency responders by telephone. Advice from 9-1-1 
operators was often to remain in place and await rescue. 5 

At 9:20 a.m., also in WTC 1, occupants frustrated by a lack of information reached out to authorities, 
including the Port Authority Police Desk. 

Male: Oh, hi. We're on the 39th floor here. We're not getting any messages. Can 

somebody tell us what's up? 

PAPD: Come on down. Just come on down, everybody just come on down. Get down 

the staircase. Don't take the elevator. 

Male: Is there smoke in the stair? 

PAPD: Check the stairway. We're not sure yet. 

Male: We've got... 

Male: We're in one, 39. We've got, uh... 

PAPD: You should be low enough to get down, because it happened up high. 

Male: All right. 

Male: Thank you. (PANYNJ 200 la) 



SI 



New York City 911 Emergency Call Recordings, 2001. 
NIST NCSTAR 1-7, WTC Investigation 1 07 



Chapter 8 



As the two communication records exemplify, calls to PAPD from occupants, in general, resulted in 
clearer, more specific information, such as where the emergency was. Further, some PAPD operators 
quickly recommended to occupants that they should leave the building, including WTC 2 prior to 9:02:59 
a.m., although that was not universally true. FDNY dispatch and 9-1-1 operators generally advised 
occupants to shelter in place and await rescue, consistent with existing protocols, but not the preferable 
course of action on September 11, 2001. 

These quotes and communication records point out several issues that occupants had to deal with during 
their evacuation. The environment in both buildings continued to worsen with time. In addition to the 
physical exertion required to descend the stairways, occupants were also faced with the need to allow 
injured persons to pass and firefighters and other emergency responders to travel up the stairs to respond 
to the fires. In addition, information from co-workers and managers at times conflicted with local and 
building-wide announcements about the event. These issues are discussed in more detail in Chapter 10. 

Occupants of both buildings faced a number of challenges in their attempts to evacuate once the planes hit 
both buildings. As building authorities and emergency responders worked to organize the evacuation 
under rapidly changing conditions, many occupants were directed out of the stairwells during their 
evacuation. Thirty-four percent of respondents to the telephone survey reported leaving a stairwell at 
least once during their evacuation. While this occurred from floors 3 through 78 in WTC 1, the largest 
percentages were found in the skylobbies on floor 78 (7 percent of telephone interview respondents) and 
floor 44 (9 percent), as well as on floor 13 (9 percent). Similarly, in WTC 2, the skylobbies on 78 
(13 percent of respondents) and 44 (25 percent) were mentioned most often. Some occupants left 
stairwells at the skylobbies or other locations to seek additional information or to find a better route out of 
the building. A survivor from a floor in the 90s in WTC 1 used multiple stairs during egress: 

"The staircase became crowded at about the 68th floor and I hoped to 
find a less crowded stairwell. I descended the stairs from 78 to about the 
50s, before exiting the stairs again. The stairs at 68 became more 
crowded because they were evacuating some injured or handicapped 
people. The speed slowed to a halt— so we exited to find another 
staircase. I entered staircase B (in the 50s), and descended to the 40s 
before exiting the stairs again to find better conditions as I exited the 
building." Interview 1000052 (NIST 2004) 

Occupants were also directed to leave the stairwells. 

A survivor who began on a floor in the 80s of WTC 1 : "At the skylobby, 
we were re-directed by Port Authority personnel. We went to the 
stairwell on the far west side of the building. I saw many people trying to 
get into the little door of the stairwell. I went back to a different stairwell 
that was empty. The center stairwell was the one we ended up in that 
took us to the ground floor lobby." Interview 1000535 (NIST 2004) 

In WTC 1, occupants with mobility impairments presented a particular challenge. A specific floor was 
designated by emergency responders to hold mobility-impaired occupants. The idea was to allow fully 
ambulatory occupants to egress, and then move the mobility-impaired occupants out once the stairwells 
had cleared somewhat. Numerous interview respondents indicated that a relatively small number of slow- 
moving occupants hindered the flow of faster moving occupants behind them in the stairwells. Reports of 



108 NIST NCSTAR 1-7, WTC Investigation 



September 11, 2001, 9:02:59 a.m. - 9:58:59 a.m., 56 Critical Minutes 



the specific floor number in WTC 1 varied (12, 18, or 20), but several survivors recalled the existence of a 
holding floor. 

A wheelchair-user from a floor in the 60s was being assisted by four 
previously unknown occupants down the stairwells in WTC 1 : "We saw 
a Port Authority person. He told us to go to the 18th floor [because] 
there was an evacuation station for people who needed special assistance. 
I was carried mostly by the two [occupants] who wandered onto my floor 
off the stairs [in the 60s], but then there were two more that also helped. 
They switched off. At about the 30th floor, we encountered firemen for 
the first time. We moved to the right." Focus Group #3 (NIST 2004) 

NIST found no evidence that a similar holding floor for mobility-impaired occupants existed in WTC 2. 



8.3 



EVACUATION ROUTE THROUGH CONCOURSE 



The first firefighter killed at the WTC was reportedly killed by a falling person (FDNY CD 12/15 2001); 
many evacuees and bystanders were injured by debris. Figure 8-1 shows the WTC plaza from the 
WTC 1 Mezzanine Level, perhaps viewed by thousands of occupants as they prepared to walk down the 
escalators to the Concourse Level. Throughout the evacuation, the lobbies and particularly the Concourse 

area served as the primary 
route to safety for survivors 
of the WTC attacks. Within 
approximately 1 minutes of 
the attack on WTC 1, 
building officials and police 
in WTC 1 and WTC 2 had 
established a procedure to 
route evacuees from both 
buildings through the 
Concourse (mall) rather than 
the plaza in order to prevent 
occupants from being 
injured by falling debris or 
people or being upset by 
shocking sights. As 
presented in Section 2.2.2, 
Stairwells A and C in both 
WTC 1 and WTC 2 
discharged to the Mezzanine 
Level, while Stairwell B 

discharged to the Concourse Level, without the choice of exiting at the Mezzanine Level. Occupants who 
used Stairwell A or Stairwell C were directed to use an escalator (see Figure 8-2) in order to descend to 




Figure 8-1. View from mezzanine level in WTC 1 looking east 

across WTC plaza, covered with debris. 'Sphere,' artwork by 

Fritz Koening, can be seen on the plaza. 



NISTNCSTAR 1-7, WTC Investigation 



109 



Chapter 8 



the Concourse Level. Many face-to-face 
interviewees who used Stairwell A or C 
reported that the escalators were turned off 
in order to allow occupants to descend 
easily using both escalators. 52 

Once through the underground shopping 
mall, an occupant originating from a floor 
in the 30s in WTC 2, recalled being 
directed out of the Concourse at 
approximately 9:03: "The security guards 
were like a human chain telling us which 
direction to go. We followed the security 
guards direction, half walking, half running 
towards Borders [retail book store] to now 
go up from the Concourse to get out." 
Interview 1000842 (NIST 2004) Ascending 
the escalator from the shopping complex 
was captured in Figure 8-3. 




Source: John Labriola 2001, edited by NIST. 

Figure 8-2. Escalator from mezzanine to 
concourse level in WTC 1 on September 11, 2001. 



8.4 MORE ANNOUNCEMENTS 
IN WTC 2 

At 9:12 a.m., the fire command station in 
WTC 2 (310-B) used WTC Ch. 22 to 
broadcast that the fire (floor) warden 
phones in WTC 2 were not working. 53 
Further, WTC personnel indicated that 
they were making announcements that 
wardens should not wait for further 
instructions over the floor warden phones, 
as per protocol. 



Figure 8-3. Evacuees leaving complex near WTC 5. 

310-B to any units: Be advised that Building Two (inaudible) only 
(inaudible) warden phones. We can't pick up warden phones. We are just 
making straight announcements telling the people not to stay at the 
warden phones, because we can't pick them up. (PANYNJ 2001a) 

NIST found no other evidence to confirm or refute that the fire command station issued such an 
announcement or that any occupant heard an announcement around that time. 




Analysis of Face-to-face Interviews (NIST 2004). 



For an explanation of the emergency communication channel assignments, see NIST NCSTAR 1-8. 



110 



NIST NCSTAR 1-7, WTC Investigation 



September 11, 2001, 9:02:59 a.m. - 9:58:59 a.m., 56 Critical Minutes 



At 9:20:10 a.m., approximately 17 minutes after the second tower was hit, the fire command station made 
another announcement over the public address system updating the occupants. The announcement was 
recorded in the background of multiple 9-1-1 calls. 54 To the best of NIST's ability to understand, the 
announcement went as follows: 

"This is the fire command station in building number two. We are in the 
lobby. We are holding down here. The condition seems to have 
subsided somewhat. There are people in the building. There are some 
people have left. If you wish to leave, you can now use the Concourse. 
You may walk to the Concourse on this side, avoiding building number 
one. The condition seems to have subsided in the lobby of building two. 
Some people have left the building. We are here monitoring the 
situation. The fire department are concentrating their efforts on building 
number one. People are leaving the building... 

(unintelligible). . .Number 2 . . . (unintelligible) ..." 

The announcement lasted between 60 and 70 seconds, with the unintelligible portion comprising 
approximately the final 10 seconds. The content of the 9:20 a.m. WTC 2 public address system 
announcement varied significantly from the pre-planned building evacuation announcement covered in 
Section 2.2.4. However, no survivors from whomNIST collected first-person accounts reported hearing 
the 9:20 a.m. announcement in WTC 2. This may be attributable to the fact that approximately 75 percent 
of all WTC 2 survivors had left WTC 2 by 9:20. Two occupants, trapped above the floors of impact, 
however, did respond to the 9:20 a.m. announcement, according to 9-1-1 records. 55 

8.5 CONDITIONS CONTINUE TO DETERIORATE IN BOTH TOWERS 

As the events unfolded, conditions within both towers continued to worsen. At or above the floors where 
the airplanes hit the buildings, occupants had to cope with fire, smoke, and severe building damage. In the 
time period after 9:15 a.m., only four calls to 9-1-1 came directly from inside WTC 1. As Figure 8-4 
shows however, many occupants above the impact region in WTC 1 remained alive at 9:58:59 a.m., 
shortly before WTC 2 collapsed. 

Below the impact floors, occupants worked to evacuate the building while emergency responders made 
efforts to organize the evacuation and gain access to damaged floors to rescue injured or trapped 
occupants. At 9:35 a.m., a WTC Official in the security command center (SCC) reported to S2 trouble 
unlocking the remotely-controlled doors: 

MaleB:S2toSCC. 

Female A: Go, S2. 

Male B : How are you doing up there? 

Female A: We've got a lot of smoke. 

Male B: We're on 16 now. 

Female A: That's a copy. We can't use software right now to try to release the doors. 

(PANYNJ2001a) 



54 New York City 9 1 1 Emergency Call Recordings, 200 1 . 
New York City 911 Emergency Call Recordings, 200 1 . 



NIST NCSTAR 1-7, WTC Investigation 1 1 1 



Chapter 8 




Figure 8-4. Occupants seeking fresh air on floors 103 - 105 on north face of 

WTC1 at 9:58:12 a.m. 

The reported difficulty unlocking doors did not appear to directly prevent any occupants from 
successfully evacuating, however, as every fourth door was always open and the doors were only locked 



112 



NIST NCSTAR 1-7, WTC Investigation 



September 11, 2001, 9:02:59 a.m. - 9:58:59 a.m., 56 Critical Minutes 



from the stairwell side. Further, only the mechanical rooms and roof were controlled electronically. The 
communication does reflect the condition of the SCC and the difficulty the conditions posed for 
operations originating from that office. Numerous other reports of fire, smoke, and damage were noted in 
transcripts of PANYNJ radio calls regarding the 22nd floor command center (PANYNJ 2001a). 

In WTC 1 , more than half of the survivors noticed smoke in their immediate area during their evacuation 
compared to only 10 percent who reported noticing smoke at first awareness. At least 25 percent reported 
collapsed walls during their evacuation compared to only 6 percent who noticed collapsed walls at first 
awareness. 



Table 8-3 shows a summary of observations by the survivors during their evacuation from WTC 1 and 
WTC 2. All percentages increased from earlier observations on their floors prior to beginning evacuation. 

Table 8-3. Observations of conditions during evacuation of WTC 1 and WTC 2. 



Observation 


WTC 1 (n=440) 


WTC 2 (n=363) 


Smoke 


57% 


21 % 


Sprinklers/water 


55% 


3% 


Fatally injured people 


41 % 


8% 


Power outage/flickering lights 


37% 


29% 


Jet fuel 


31 % 


21 % 


Fallen ceiling tiles 


29% 


12% 


Fire alarms 


26% 


19% 


Collapsed walls 


25% 


11 % 


Extreme heat 


10% 


7% 


Fire 


9% 


4% 


Fireballs 


3% 


2% 



Conditions in the stairwell were challenging, as well. 

A survivor from a floor in the 20s in WTC 1 : "The stairwell was lit the 
entire way down. There was a grayish color smoke which smelled like 
fuel. The more we reached the lower floors the stronger the smell 
became. On the 6th floor, the sprinklers were on, which slowed us down 
because we wanted to be cautious and not slip or fall." Interview 
1000044 (NIST 2004) 

An occupant who survived the horror of aircraft impact on the 78th floor 
in WTC 2:"... not much debris, but there was a pipe that we had to go 
under, hanging wire, and water in the stairs; no counter flow after we 
were hit, stairs were well lit, but 78 had no lighting. On 40-well lit, a lot 
of doors were locked, stranger smell that I can't identify. As we were 
walking down, we saw fireman coming up and told us to get to 40 and 
that someone would take us in an elevator down to the lobby." Interview 
1000562 (NIST 2004) Firefighters and security personnel would escort 



NISTNCSTAR 1-7, WTC Investigation 



113 



Chapter 8 



the occupant down in the elevator, out through the Concourse area to a 
waiting ambulance. 

At 9:36 a.m., an occupant's telephone call to a New York City 9-1-1 telephone operator indicated that a 
floor in the 90s of WTC 2 had collapsed. According to NYPD records, information from this call 
concerning the floor collapse appears to have been conveyed inaccurately by the 9-1-1 call taker and the 
NYPD radio dispatcher. 56 The NYPD dispatcher transmitted the message at 9:41 a.m. and 9:51 a.m. 
Independent of the exact floor number, the call content demonstrated deteriorating conditions above the 
impact region in WTC 2. 

Table 8-4 shows the total evacuation time for survivors in WTC 2. Total evacuation time was measured 
from first awareness to exiting the building and included time spent on the starting floor. While the total 
evacuation time increased from the lower floors to the middle floors, as would be expected when the 
proportion of occupants who used elevators was roughly constant, the total evacuation time for survivors 
from the upper floors decreased due to the increased proportion (approximately twice as many, see 
Section 6.2.3) of survivors who used elevators to evacuate. 

Table 8-4. Total evacuation time (min) for survivors from WTC 2. 



Total Evacuation Time" (min) 


25% 
Evacuated 


50% 
Evacuated 


75% 
Evacuated 


Average Time 
(min) 


Lower floors (Basement - 42) 


10 


15 


21 


17.6 


Middle floors (43-76) 


23 


35 


45 


34.9 


Upper floors (77-110) 


11 


23 


36 


25.5 


All survivors in WTC 2 


13 


21 


35 


25.0 



a. Total evacuation time was measured from first awareness to exiting the building. 

Note: Italics indicate the presence of a bias in the underlying data. Interpretation and comparison of data should consider 

this bias. 

Source: NIST WTC telephone survey data. 

Minutes prior to the collapse of WTC 2, an NYPD Emergency Services Unit (ESU) officer radioed from a 
floor in the 20s to the outside that he was having trouble ascending the stairwell due to the large number 
of occupants descending (Interview 24 NYPD [NIST 2004]). While the origin of the occupants remains 
unknown, only 1 1 occupants who started evacuating below the impact region were known not to have 
survived. 

WTC 2 collapsed at 9:58:59 a.m., leaving over 600 occupants and over 100 emergency responders dead. 



NYPD, McKinsey and Company, NYPD call-routing and message dispatch, "106' floor is collapsing." Draft summary report 
July 23, 2002. 



114 



NIST NCSTAR 1-7, WTC Investigation 



Chapter 9 

September 11, 2001, 9:58:59 a.m. - 10:28:22 a.m. 

Collapse 



"The lights were off and it was dark. You could barely see your hand in 
front of your face and there was a lot of dust. We just stood there and 
waited for the dust to settle. The emergency lights went on and we tried 
resuming going back down those stairs for half a flight. People [were 
coming back] up the stairs and said, 'the egress is gone.'" Interview 
1000594 (NIST 2004) 

At 9:58:59 a.m., many of the 624 to 680 occupants and building personnel who had fought for survival in 
WTC 2 for the 56 minutes after the building was attacked were killed by the building collapse. Over 100 
emergency responders, unaware of impending collapse, were killed as well (NIST NCSTAR 1-8 2005). 
The effect of the collapse of WTC 2 was felt not only at the remaining WTC buildings, but all over Lower 
Manhattan, as shown in Figure 9-1. 

9.1 EVACUATING WTC 1 

Although occupants in WTC 1 may or may not have comprehended the magnitude of the collapse at the 
time, it still had a significant impact on the evacuation of remaining survivors in WTC 1 . As described 
below, survivors related to NIST, during face-to-face interviews, the nature of the cloud of debris and 
smoke which filled the Concourse and lower portions of the stairwells in WTC 1 , making evacuation that 
much more difficult for those still left in the building. 

A survivor in the shopping Concourse who began on a floor in the 50s of 
WTC 1 : "We heard (what we later learned) the South Tower/WTC 2 
collapsing. We stopped near the joining of the Gap store and the PATH 
escalators. The sound, at first, was muted and rumbling. It was a 
scraping sound that grew louder over the course of 15-30 seconds. It 
seemed like it was going to envelop us. The sound became very loud 
(like steel on steel). Very quickly after WTC 2 collapsed, the lights went 
out in the Concourse. Immediately, people screamed and panicked, and 
started to run in the dark. This was the one time I thought I really was 
going to die. I saw what I thought then was a wall of water, like a 
tsunami, coming at me (I thought I was going to drown). What it turned 
out to be was the cloud of debris and smoke. It irritated my eyes and 
filled my mouth, and made me cough and spit. I called to the others (my 
staff/ co-workers) who were with me. I felt the need to be in touch with 
them, since we had come all this way together." Interview 1000106 
(NIST 2004) 



NIST NCSTAR 1-7, WTC Investigation 1 15 



Chapter 9 




©2001. N ew York C i ly Police De pa r t m ent. All rights reserved. 

Figure 9-1. Collapse of WTC 2, viewed from north. 

A survivor still in a stairwell who began on a floor in the 30s of WTC 1: 
"We descended on this stair until we got to the landing between floors 7 
and 8. The lights now went out in the stairwell. I entered the 8th floor 
and went down a corridor to a different staircase (I'm not sure which 
stair, but I believe it was the original one I started out on). People on the 
stairs below me (who apparently got a big rush of air) said the floors 
below us were collapsing. They turned around and started running back 
up and they were covered with a gray, powdery substance, such as ash or 
smoke — so we believed them. Then when the fireman opened the door 
on the 8th floor and said to come onto the 8th floor, I readily followed 



116 



NIST NCSTAR 1-7, WTC Investigation 



September 11, 2001, 9:58:59 a.m. - 10:28:22 a.m., Collapse 



his instructions. Some didn't follow those instructions and continued 
back upstairs. I descended the stairs from the 8th floor to what I believe 
was the 5th floor. It was still dark in the stairwell. I changed stairwells 
for the last time and descended to the Mezzanine Level. I exited the stairs 
onto the Mezzanine, headed toward the west wall, then the north wall, 
and finally exited through the emergency exit door." Interview 1000042 
(NIST2004) 

A survivor from a floor in the 50s in WTC 1 : "We thought it was a bomb 
at this point, however, it was Building Two collapsing. Vibrations 
stopped; the noise was incredible; the lights went out and the sprinklers 
came on. We walked a lot more quickly. We were getting out of the 
building. It was hard to walk down the stairs because of the water 
coming from above from the sprinkler system. There was a large amount 
of smoke and I couldn't see the person in front of me. I guess we smelled 
jet fuel. I could barely see the emergency strips on the stairs, but you 
could see them [because] the lights had gone out." Interview 1000054 
(NIST 2004) 

A survivor in the stairwell who began in the teens of WTC 1 : "We heard 
a gigantic explosion. We reached the second floor landing and we 
stopped. It was pitch black; black and tremendous suffocating smoke and 
dust. We stopped there and choked and froze. The lights went back on 
after it seemed like forever. We were told by the firemen and rescue 
workers that we had to go up to the fourth floor, that the exit was 
blocked." Interview 1000093 (NIST 2004) 

Whereas occupants were primarily using the cover of the underground Concourse Level to escape the 
WTC site prior to the collapse of WTC 2, the debris created by the collapse forced the remaining 
occupants to proceed outdoors after they left the stairwells. For occupants leaving Stairwells A and C on 
the Mezzanine Level, the exit on the North wall, West corner provided access to the Tobin Plaza, from 
which many proceeded North between WTC 5 and WTC 6; others used the bridge to cross West St to the 
World Financial Center. Occupants leaving Stairwell B on the Concourse Level moved West to the West 
Street VIP entrance, where many continued to the World Financial Center, while others traveled North 
along West St. 

The evacuation rate slowed dramatically in WTC 1 after WTC 2 collapsed. Clearly, the obstacles 
presented by the dust, debris, and loss of power from the WTC 2 collapse contributed to slowing the 
occupants travel speed. By 9:58:59 a.m., 88 percent of survivors (over 6,500 people) in WTC 1 had 
exited the building. Approximately 900 additional occupants and many emergency responders would exit 
WTC 1 over the remaining 29 minutes before the building collapsed. 

Many mobility-impaired occupants remained in a holding pattern on a lower floor in WTC 1 during this 
time. An FDNY Battalion Chief wrote "Last Man Down," in which he described walking onto the 12th 
floor of WTC 1: 

"The office was filled with people. Forty, fifty, sixty of them. I didn't 
stop to count, but there they were, in all shapes and sizes, of all ages. All 
of them were sitting quietly, patiently, apparently waiting for instruction 



NIST NCSTAR 1-7, WTC Investigation 1 17 



Chapter 9 



or assistance... There had to have been a couple dozen nonambulatory 
office workers in that room, easy, and each one of them had one or two 
or sometimes three friends helping them. Friends, coworkers, people 
they'd just met in the stairs, for all I knew." (Picciotto 2002) 

Some occupants, however, had resisted the idea of delaying their evacuation by stopping on the 
designated holding floor. A mobility-impaired occupant who started in the 60s had been instructed while 
in the stairwell by a Port Authority employee to report to the 1 8th floor, when the time came to choose 
between resting or continuing to evacuate: 

"At about the 22nd floor, Building Two collapsed. The building swayed. 
There was smoke and dust from the collapse of building two. I had my 
mask on, but it was hard to breathe and see. We picked up speed. 
Firemen started coming back down (from above us). We had to leave the 
staircase because we couldn't see or breathe. We went into a tenant's 
space for 5 or 10 minutes. Then, we couldn't find another staircase so 
we went back to the one we had just left. When we got to the 1 8th floor, 
we just kept going. The [helpers] asked me if I wanted to stay, I said 
"No." I had been with them this far. After that we were moving more 
quickly." [NIST Focus Group #3, 2004] 

Battalion Chief Picciotto described ordering the evacuation of the holding floor in "Last Man Down," 
largely steering them down Stairwell B, at approximately 10:21 a.m. The fate of those mobility-impaired 
occupants remains unclear. The conditions inside the stairwells, after WTC 2 collapsed, became more 
difficult. 

A survivor who began on a floor in the 60s of WTC 1 : "The staircase 
was a mess; it was filled with concrete, pieces of metal sticking out of 
the stairs. There were also holes on the walls. There was also lots of 
water on the floor." Interview 1000543 (NIST 2004) 

While building personnel attempted to provide building-wide announcements in WTC 1 , only two 
survivors remembered hearing announcements in WTC 1, both near the time of the collapse of WTC 2. 
Specific details of the announcements were not recalled beyond "some sort of an emergency" or "to 
evacuate the building" Interviews 1000594 and 1000068 (NIST 2004), respectively. 

After WTC 2 collapsed, there was a sharp decrease in 9- 1 - 1 call volume from inside WTC 1 . Above the 
floors of impact in WTC 1, NIST found only one call to 9-1-1 from a person inside WTC 1. 

Table 9-1 shows the total evacuation time for survivors in WTC 1 overall, as well as by zone (low, 
middle, and high). Comparisons of upper zone data with data from lower and middle zones should be 
carefully considered, however, as the upper zone represents a smaller region (fewer number of floors and 
less vertical height represented) than the lower and middle zones. This was due to the fact that no 
individuals survived from above the 9 1 st floor. 



Several calls were received from persons relaying information on behalf of occupants in WTC 1, but the time delay in relaying 
the information from the occupant to the 9-1-1 dispatcher was unknown. 

118 NIST NCSTAR 1 - 7, WTC Investigation 



September 11, 2001, 9:58:59 a.m. - 10:28:22 a.m., Collapse 



The quartile and mean times for survivors to exit the building shown in Table 9-1 increased as the 
survivors were higher in the building. The first quartile of surviving evacuees from the middle zone were 
exiting WTC 1 at approximately the same time as the last quartile of surviving evacuees from the lower 
zone (approximately 40 minutes). Similarly, the first quartile of surviving evacuees from the upper zone 
were exiting WTC 1 at approximately the same time as the last quartile of surviving evacuees from the 
middle zone (approximately 60 minutes). Note, however, that the number of floors (and vertical range) of 
the upper zone was significantly less that the lower and middle zones. Therefore, comparisons of the 
upper zone times to the lower and middle zone times should factor the zone differential into the analysis. 



Table 9-1. Total evacuation time (min) for survivors 1 


from WTC 1. 


Total Evacuation Time" 
(min) 


25% 
Evacuated 


50% 
Evacuated 


75% 
Evacuated 


Average Time 
(min) 


Lower floors (Basement - 42) 


16 


27 


42 


29.0 


Middle floors (43-76) 


41 


51 


65 


54.7 


Upper floors (77-91) 


60 


71 


77 


70.3 


All survivors in WTC 1 


25 


41 


58 


41.9 



a. Total evacuation time was measured from first awareness to exiting the building. 

Note: Italics indicate the presence of a bias in the underlying data. Interpretation and comparison of data should consider 

this bias. 

Source: NIST WTC Telephone Survey Data. 



9.2 



LOCATION OF VICTIMS AND SURVIVORS WITHIN THE BUILDING 



Once the two towers collapsed, the fate of 2,749 building occupants and emergency responders was 
sealed. Table 9-2 shows the likely location at time of impact of all decedents identified by the City of 
New York as decedents. It also categorizes types of emergency responders who perished in the disaster. 
To identify locations, NIST relied on more than 300 face-to-face interviews and 800 telephone surveys, 
various Web sites maintained by survivors or victims' families and colleagues, several media outlets' 
reports, and a badge list maintained by The Port Authority of New York and New Jersey. Table 9-2 
below shows the results of that analysis. 

The following sources were used to identify the likely location of decedents at time of impact: 

• Septemberl lVictims.com: This site is dedicated to the victims of September 11, 2001 tragedy. 

• Portraits: 9/1 1/01 : Published by the New York Times in 2003, this book includes short interviews 
with family members of many decedents. 

• CNN.com In-Depth Special (http://www.cnn.com/SPECIALS/2001/memorial/index.html): 
Tribute site for people to write remembrances of decedents. 

• Badge List maintained by Port Authority of New York and New Jersey: Includes name, 
employer, building, and floor for all occupants with badge-access to WTC 1 or WTC 2. 

• Numerous memorial sites maintained by companies which lost employees: Includes names and 
remembrances of decedents. Examples include the Port Authority, Fire and Police Departments, 
Marsh & McLennan Companies, EuroBrokers, Fiduciary Trust, and others. 



NISTNCSTAR 1-7, WTC Investigation 



119 



Chapter 9 



• Newsday.com: Includes short stories written about specific decedents. 

Table 9-2. World Trade Center decedent location. 



Likely Location at Time of Impact 3 




World Trade Center 1 Occupants 


1,462 


At or Above Impact 


1,355 


Below Impact 


107 


World Trade Center 2 Occupants 


630 


At or Above Impact 


619 


Below Impact 


11 


Confirmed Below Impact in WTC 1 or WTC 2 


30 


Unknown Location Inside WTC 1 or WTC 2 


24 


Emergency responders (Total) 


421 


FDNY 


343 


NYPD 


23 


PAPD 


37 


Hospital/Paramedic 


7 


Federal 


2 


Volunteer Responders 


9 


Bystander / Nearby Building Occupant 


18 


American Flight 11 


87 


United Flight 175 


60 


No Information 


17 


Total 


2,749 



a. Where possible, NIST has used eyewitness accounts to place individuals. Where no specific 
accounts existed, NIST used employer and floor information to place individuals. 

The analysis shown in Table 9-2 generally assumes that individuals were on their work floor when his or 
her building was hit. The employer and floor number from the badge list were generally used to 
designate location. An employee of Cantor Fitzgerald, for example, would be assumed to have died 
above the floors of impact in WTC 1, unless evidence to the contrary was discovered. Often, information 
about the activities or location of an individual was available from on of the sources described above, and 
that information was used to categorize individuals. Additional information was necessary to place the 18 
individuals documented as bystanders or occupants of nearby buildings. There were 30 individuals who 
were known to have been below the floors of impact, but whose locations could not be determined. 
Typically, these individuals were security guards and fire safety staff who were observed performing 
activities below the floors of impact after the airplanes struck. There were 24 individuals who were likely 
in the building, but for whom no clarifying information could be uncovered indicating whether they were 
above or below floors of impact. Their occupations largely included maintenance, janitor, delivery, 
safety, or security functions. 



120 



NIST NCSTAR 1-7, WTC Investigation 



September 11, 2001, 9:58:59 a.m. - 10:28:22 a.m., Collapse 



Emergency responders were defined to be people who arrived at the site from another location; thus, 
security staff and Port Authority staff (different from PA Police Officers) were not defined as emergency 
responders. 

No information could be determined for 17 individuals. 

Estimated Number of Decedents below the Impact Regions 

The total number of people who perished below the floors of impact was: 

421 emergency responders; 

107 individuals in WTC 1; 

1 1 individuals in WTC 2; 

30 individuals below the impact region in either WTC 1 or WTC 2; 

Up to 24 individuals were somewhere in the building, but may have been above or below impact 
regions in either building; 

Up to 17 individuals for whom no information could be located; 

Thus, there were between 569 - 610 individuals who were below the impact floors but did not 
successfully evacuate. Assuming that all of these individuals were alive at the time that his or her 
respective building collapsed, approximately 20 percent of those lost on September 11, may have been 
alive just prior to collapse of the WTC towers. 

Estimated Number of Decedents at or above the Impact Region 

The total number of individuals who perished who were at or above the floors of impact was: 

• 1,355 individuals above the 91st floor in WTC 1 at 8:46:30 a.m.; 

• 619 individuals above the 77th floor in WTC 2 at 9:02:59 a.m.; 

• Up to 24 individuals were somewhere in the building, but may have been above or below impact 
regions in either building; 

• Up to 17 individuals for whom no information could be located. 

Therefore, between 1,974 and 2,005 individuals were at or above the area of impact in WTC 1 and 
WTC 2 who did not successfully evacuate. 



NIST NCSTAR 1-7, WTC Investigation 121 



Chapter 9 



Number of Aircraft Passengers and Crew 

The total number of airplane passengers and crew was 147, not including the 10 hijackers. 58 American 
Flight 1 1 had 87 passengers and crew, while United Flight 175 had 60 passengers and crew. 

Other Victims 

Additionally there were 18 people (non-WTC occupants) killed outside the building by debris and jet fuel. 



9.2.1 



Fatalities in Elevators 



USA Today (Cauchon 2001) reported that as many as 200 occupants died in elevators in WTC 1 and 
WTC 2. While the NIST Investigation did not produce an estimate of the number of occupants who died 
in elevators, NIST did collect information related to elevators and elevator use that may be of benefit in 
identifying ways of reducing building occupant risk in general. 

The elevators were the normal means of vertical travel in WTC 1 and WTC 2, and it was likely that many, 
if not most, of the elevators were in use at the time of aircraft impact. There were two ways in which 
lives of building occupants may have been lost on September 11, 2001 while using the elevators. First, if 
the cables above the elevator car were severed, the car would have dropped to the bottom of the shaft 
(which may not be the basement, depending upon the specific elevator shaft). The safety brakes on the 
elevator car would have been activated by the governor cable when the governor wheel exceeded the 
design speed. Since the governor was mounted at the top of the hoistway, should the governor cable have 
been cut along with the hoist cables, the safety brake would not have been activated, and the car would 
have fallen. If the car dropped far enough, occupants of the car may not have survived the impact. For 
shorter falls, the occupants may have survived but remained trapped in the elevator car at the bottom of 
the shaft. 



Each of the towers had 99 elevators serving the above ground floors, but most were locals that ran to only 
a few floors within one of the three elevator zones. For an elevator's cables to be cut and result in 
dropping the car to the bottom of the shaft, the cables would need to have been in the aircraft impact 
debris path, floors 93 through 98 in WTC 1 or floors 78 through 83 in WTC 2. Inspection of the elevator 
riser diagram and architectural floor plans for WTC 1 shows that the following elevators met these 
criteria: cars 81 through 86 (Bank B) and 87 through 92 (Bank C), local cars in Zone III; car 50, the 
freight elevator, and car 6, the Zone III shuttle. Inspection of the elevator riser diagram and architectural 
floor plans for WTC 2 shows that the following elevators met these criteria: cars 75 through 80 (Bank A), 
local cars in Zone III; car 50, the freight elevator, and car 6, the Zone III shuttle. Bank A local cars could 
have fallen a maximum of 6 floors, Bank B cars 16 floors, and Bank C cars 23 floors. Cars 6 and 50 
could have fallen all the way to the pit in the sub-basement level, and car 50 in WTC 1 was reported to 
have done so. 59 Elevator 81 (Bank B) in WTC 1 was apparently not destroyed, as WTC Radio Channel 
25 (maintenance and electric) recorded a request for assistance in getting the elevator doors opened and 
noted smoke and water infiltration into the elevator cab (PANYNJ 2004). 



New York City has not issued death certificates for any hijackers. 
' NIST Interview, 2004. 



122 



NIST NCSTAR 1-7, WTC Investigation 



September 11, 2001, 9:58:59 a.m. - 10:28:22 a.m., Collapse 



In WTC 1, several occupants seeking to use elevators reported that the elevators were not working 
immediately after impact, despite signage and a training program specifically prohibiting elevator usage 
in the event of an emergency (see Chapter 4). In WTC 2, analysis of telephone interviews determined 
that approximately 1 8 percent of the survivors evacuated primarily using an elevator on 
September 11, 2001. As with WTC 1, however, it remains unknown how many occupants were using the 
elevators at the time of impact. In addition to the occupants who may have been killed while using 
elevators, several occupants reported numerous fatalities in the 78th floor skylobby, where occupants 
were either waiting for elevators or deciding whether to return to their offices when the second aircraft 
struck WTC 2. Interview 1000562 (NIST 2004) and (NIST NCSTAR 1-7A). 

Not all occupants in an elevator when their building was hit were killed. Some occupants were able to 
force doors open, a few 'chopped' their way through walls (one group famously used a window washing 
squeegee [Dimarco 2004]), while others were able to climb out of elevators which had fallen to the 
bottom of their respective shafts (Cauchon 2001). Figure 9-2 shows occupants leaving an elevator in the 
lobby of WTC 1 shortly prior to the collapse of WTC 2 on September 11, 2001. The reason the 
occupants were able to escape after an extended entrapment remains unclear, although Port Authority 
officials speculate that the elevator may have reverted to slow-speed (50 fpm [0.25 m/s]) run, designed to 
prevent entrapment. 60 




© 2001 Goldfish Pictures 






Figure 9-2. Occupants leaving a previously stuck elevator on September 11, 2001, in 

the lobby of WTC 1. 

Second, a number of safety systems and sensors in the WTC 1 and WTC 2 elevator systems were 

arranged to stop the elevator in place if they detected an abnormal condition (PA Interview 5 

[NIST 2004]). Stopping an elevator car generally represents the safest course of action. The intercom in 



'Public Comments to Draft NCSTAR 1-7. PANYNJ. 2005. 



NIST NCSTAR 1-7, WTC Investigation 



123 



Chapter 9 



every car would then be used to notify authorities of an entrapment. Given the size of WTC 1 and WTC 2 
and the number of elevators, however, occupant extraction could have taken hours, as evidenced by the 
five or more hour waits for some occupants during the 1993 bombing (Isner and Klem 2003). 

In fires and earthquakes, however, entrapment may result in additional risks to the occupants before they 
can be removed. In an earthquake, for example, one of the greatest safety risks comes from the 
possibility that the counterweight jumps its track and is struck by the car. Seismic switches on elevators 
in seismic zones are designed to detect the earthquake and cause the system to stop, move the car in the 
direction away from the counterweight to the next floor, stop and open the landing doors (ASME 17.1, 
Safety Code for Elevators and Escalators). 

Another issue regarding entrapment was the presence of "door restrictors." A door restrictor is a safety 
device required on all new elevators (ASME 17.1, Safety Code for Elevators and Escalators) and retrofit 
on existing elevators (ASME 17.3, Safety Code for Existing Elevators) upon any elevator equipment 
upgrade. Door restrictors mechanically prevent the elevator car doors from being forced open unless the 
car floor is within 36 in. of a landing floor. They were installed in some, but not all, elevators in WTC 1 
and WTC 2 (Interview PA 5 2005). Door restrictors address the significant safety hazard that occurs 
when occupants force the car doors open and fall through the gap down the shaft or get crushed between 
the car and shaft wall. Door restrictors have all but eliminated these accidents, but have made self-escape 
by entrapped occupants nearly impossible. Elimination of door restrictors alone would lead to increased 
deaths from accidents. A better solution would address both issues. 

Occupants could have become entrapped at any level if the aircraft impact caused the elevators to stop. If 
the car floor were not within 36 in. (plus or minus 18 in.) of a landing floor, the door restrictor would 
have prevented the car doors from being forced open from the inside, and the roof hatch could only be 
opened from the top of the car, making for self-escape extremely difficult. If the car was within 36 in. of 
a landing, occupants could have forced the car doors open and then would have had to force open the 
landing doors or break through the elevator shaft wall, which was 3 in. to 3.5 in. of gypsum material. 



124 NIST NCSTAR 1-7, WTC Investigation 



Chapter 10 

Discussion and Analysis 



The attacks on the World Trade Center (WTC) towers were unprecedented events, extending well beyond 
the typical engineering design scenarios of buildings on many levels. Occupants were forced to deal with 
a full evacuation of severely damaged buildings, although training and procedures called only for partial 
evacuation near affected floors. Emergency responders and building management were forced to deal 
with multiple extreme disasters in the buildings while facilitating the safe egress of some occupants and 
attempting the rescue of others. 

This chapter provides a discussion of the factors that affected the overall safety of occupants on 
September 11, 2001. It is divided into several topics areas: a summary of total evacuation time for 
WTC 1 and WTC 2 from the telephone survey data, an analysis of the important factors that affected total 
evacuation time and its components of delay prior to evacuation, and evacuation time in stairwells or 
elevators. The face-to-face interview data also highlight additional factors that were not included in the 
telephone survey, but were prevalent in sufficient number in the face-to-face data to warrant discussion. 

10.1 OVERVIEW OF TOTAL EVACUATION TIME 

Figure 10-1 shows the percentage of survivors who reported evacuating over five minute intervals, 
measured from the moment that WTC 1 was attacked. There was a large peak in evacuation in WTC 2 
during the time period immediately surrounding the WTC 2 impact. Recall that WTC 2 was attacked at 
9:02:59 a.m., approximately 16 minutes after WTC 1 was attacked. Nearly one in six WTC 2 survivors 
reported having left the building in a single five minute period between 9:01 a.m. and 9:06 a.m. Over 
40 percent of the surviving occupants left WTC 2 prior to 9:02:59 a.m. That so many evacuated was 
largely due to occupant use of the elevators in WTC 2 to evacuate prior to impact. Approximately one in 
six surviving occupants (18 percent) initially used an elevator to evacuate WTC 2. After WTC 2, was hit, 
only one elevator was operating, under the control of the Fire Department of New York. 

Integrating the rates shown in Figure 10-1 yields Figure 10-2, which represents the fraction of survivors 
remaining in the building at different points in time. Within the bounds of statistical uncertainty, the 
same number of occupants successfully evacuated WTC 1 as WTC 2. Therefore, comparing the slopes of 
the curves is appropriate. For a period of approximately 30 minutes in WTC 2 (when only stairwells were 
available) the slope (or evacuation rate) roughly matches the slope of the first 70+ minutes of evacuation 
in WTC 1 . This result would be expected given the similarity of the egress systems in the two buildings. 
The evacuation rate declines significantly for the final twenty minutes before each building collapses, due 
to the fact that over 90 percent of the occupants who were physically able to affect their own escape had 
done so by that time. Further, the collapse of WTC 2 slowed the evacuation rate in WTC 1 due to the 
significant quantity of dust and debris complicating the path from the bottom few stair landings out and 
away from WTC 1 . 



NIST NCSTAR 1-7, WTC Investigation 125 



Chapter 10 




WTC 1 
WTC 2 



10 15 20 25 30 



09 



35 40 45 50 55 60 65 70 75 80 85 90 95 100 105 

Time Since WTC 1 Attack (min) 



Figure 10-1. Reported evacuation time for survivors of WTC 1 and WTC 2. 



§ 70% 



=> 40% 



30% 



10% 



20 



40 60 

Time from WTC 1 Attack (min) 



80 







WTC1 
<~ollapsp 


1 \\ 




► 


\ 

\ 






\ 


\ WTC 1 WTC 2 
\ Collapse 












i WTC 1 
! Attack 


\ \ 




1 < 


\ \ 

^ MITC 2 \ 












WTC 2 \ \. 
Attack V ^V 






4 ^ N 




, -—- ., 



100 



Figure 10-2. Percentage of survivors remaining in the building for WTC 1 and WTC 2. 



126 



NIST NCSTAR 1-7, WTC Investigation 



Discussion and Analysis 



While Figure 10-2 provides a comparison of overall rate of evacuation in WTC 1 and WTC 2, analyses 
should also take into account that many occupants did not survive the attacks. Recasting Figure 10-2 to 
include those who did not survive results in Figure 10-3, which shows the percentage of the total building 
population remaining in the building over time. Note that the slopes of the two curves in Figure 1 0-3 are 
not directly comparable since twice as many occupants perished in WTC 1 as did in WTC 2. When 
occupants who were unable to evacuate are included, more than 80 percent of the occupants of WTC 1 
survived and more than 90 percent of the occupants in WTC 2 survived. 






=3 

DQ 



Q. 
O 
CD 
Q. 



CD 

& 

n. 
<D 
o 

CD 
Q. 



100% 
90% 
80% 
70% 
60% 
50% 
40% 
30% 
20% 
10% 
0% 



10 





WTC 1 
Collapse 


I \x 






\ ^S. WTC 1 




M V 


WTC 2 
Collapse 






► 




iWTC 1 I \ 
j Attack V \ 




i 4 \wtc 2 ^>w 








| WTC 2 V 

| Attack -. 






*«—«">► 







20 



30 40 50 60 70 

Minutes From WTC 1 Attack 



80 



90 



100 



Figure 10-3. Percentage of occupants remaining in the building for WTC 1 and WTC 2. 

Evacuation time, however, has two primary components which contribute to the total time: the delay time 
prior to initiating evacuation and the time spent evacuating. Each of these components is discussed in 
detail below, incorporating multiple regression analysis (causal modeling) of each. 

10.1.1 Evacuation Initiation Delay Time in WTC 1 and WTC 2 

Evacuation initiation delay time for occupants of WTC 1 and WTC 2 was discussed briefly in Chapter 6. 
The analysis presented here explores whether statistically significant differences existed in the overall 
evacuation initiation times for occupants in WTC 1 and WTC 2 and among occupants located in the high 
(floor 77 and higher), middle (floors 43 to 76), and lower (basement to floor 42) regions of each tower 
(for a total of six zones). Histograms of reported delay time based upon the telephone interview results 
showed that each zone had a peak evacuation initiation around minute one or minute two, generally 
decreasing over time (or skewed right). The general distribution for all six zones was a non-normal 
distribution, approximated by a gamma or log-normal distribution. 



NISTNCSTAR 1-7, WTC Investigation 



127 



Chapter 10 



Note that due to the probable bias in the data from the high region in WTC 2 discussed previously 
(occupants who delayed a length of time sufficient to prevent them from descending below the 78th floor 
prior to 9:02:59 a.m. were disproportionately removed from the telephone interview sample), no further 
consideration was given to comparing the results from the high zone in WTC 2 for the variable of 
evacuation initiation delay time. 

In order to calculate whether differences between the remaining five zones were statistically significant, 
the reported data were transformed to better approximate a normal distribution and, thus, satisfy statistical 
testing assumptions. A Box-Cox Power analysis revealed that all five zones had peak normal probability 
plot correlation coefficients (of approximately 0.95 or better) when the Box-Cox power was 
approximately zero, which indicated a log transformation. After the reported evacuation initiation delay 
times were log-transformed, an analysis of variance (ANOVA) F-statistic was calculated for WTC 1 and 
WTC 2. 

Meaningful differences in the evacuation initiation delay time in WTC 1 among the three zones (high, 
middle, low) existed, significant at the 1 percent level. Exploring the interzonal differences in WTC 1, 
there was not a statistically significant difference in the reported evacuation initiation delay times between 
the low zone and the middle zone. The high zone, however, exhibited significantly longer reported delay 
times, when compared to both the low zone and the middle zone. The increased delay time in the high 
zone compared to the low zone was significant at the 2 percent level, while the increased delay time in the 
high zone compared to the middle zone was significant at the 1 percent level. 

For WTC 2, evacuation initiation delay time was higher in the middle zone, compared to the lower zone; 
the difference was significant at the 5 percent level. Again, comparing the high zone to either the middle 
or lower zone in WTC 2 was not considered due to the bias in the upper zone delay time results. 

In conclusion, the general trend in WTC 1 and WTC 2 was for evacuation initiation delay to increase with 
building height, although there may be other explanations for this phenomenon. In WTC 1 , occupants 
reported being trapped by debris and building damage, observing fire and smoke, and assisting injured 
colleagues at a higher rate near the impact region (high) than was reported in the lower or middle zones. 
The causal modeling in Section 10.2.1 explores this further through the independent variables 'floor' and 
'environmental cues.' 

Additionally, comparison of evacuation initiation delay times across buildings should be made with 
considerable caution. In WTC 1, the cues which led occupants to decide to evacuate were substantively 
different from those that likely influenced the decision process in WTC 2. In WTC 1, the building had 
been attacked, stairwells were the only available route out of the building, and there were generally no 
instructions delivered over the public address system. WTC 2 had not yet been attacked for over 
90 percent of the occupants by the time they initiated evacuation. Thus, for less than 10 percent of the 
WTC 2 survivors, their building had been attacked by the time they chose to initiate evacuation. In 
addition, in WTC 2 stairwells and elevators were available for most people to choose to use, and multiple 
(though conflicting) public address system announcements were made. Note, however, that many 
occupants in WTC 2 did not hear any announcements, because they had evacuated the building prior to 
the time the announcements were delivered. 



128 NIST NCSTAR 1-7, WTC Investigation 



Discussion and Analysis 



10.1.2 Time and Average Travel Speed in the Stairwells, WTC 1 

The reported time spent in the stairwells for occupants of floor 10 to floor 91 in WTC 1, based on the 
NIST telephone interviews (n=368), yielded a mean normalized travel speed of 1.3 floors per minute. 
Normalized travel speed is defined as the total time from entering the stairwell until leaving the building, 
divided by the number of floor that had to be descended. According to Figure 10-4, 25 percent of the 
occupants traveled faster than 1.5 floors per minute, while 25 percent of the occupants traveled more 
slowly than 0.9 floors per minute. The median normalized travel speed was 1.2 floors per minute. The 
data include all interruptions to the evacuation process, including crowding, transfer floors, smoke, water, 
switching stairwells, and resting. Assuming an average floor height of 3.7 m (12 ft), the distance along 
the stair slope (including landings) would have been approximately 10 m (33 ft), yielding a movement 
speed of approximately 0.2 m/s (0.65 ft/s) for the median occupant in WTC 1 while in the stairwells, 
which is on the slow end of published scientific literature values for stairwell travel speeds. This is 
understandable given the frequency of crowding and the significant number of obstacles to evacuation 
reported by many occupants. 



0.15 




1 1.5 2 

Floors per Minute 



Figure 10-4. Stairwell travel speed in WTC 1 for all stairwells. 

A "rule-of-thumb" for calculating evacuation flow rate is to assume that a standard size door at the 
bottom of the stairwell is capable of discharging approximately one person per second. (Fruin 1987) 61 By 
that logic, with three stairwells, the WTC system would have been capable of moving approximately 



Fruin reported values of 40 - 60 persons per minute for free-swinging doors. The "rule-of-thumb" captures the high end of the 
reported range. 



NISTNCSTAR 1-7, WTC Investigation 



129 



Chapter 10 



three people per second from the occupied floors to the Mezzanine or Concourse. In WTC 1, where 
elevators were not operational for the duration of the evacuation, approximately 7,500 occupants used the 
stairwells over a 100 minute period, yielding a flow rate of 1.3 people per second from the occupied 
floors. Even discounting the final 27 minutes (after the collapse of WTC 2 when the flow rate dropped 
significantly), the flow rate was 1.5 people per second (or 0.5 people per second per door), or about one- 
half the "rule-of-thumb" flow rate. This is consistent with the previous observation that the stairwell 
movement speed was slower than the published literature values would have predicted and that the slower 
rate was to be expected given the number of obstacles to egress encountered by the evacuees and the total 
travel distance required. 

Impact of Wider Stairwell on Stairwell Evacuation Times 

An analysis of the hypothesis that wider stairwells correspond to faster overall evacuation times is most 
reliable when all other variables are held constant. In WTC 1 on September 11, 2001, however, many 
other variables were not constant, complicating the comparison: (a) Stairwell B (56 in. wide) exited into 
the Concourse, while Stairwells A and C (44 in.) exited to the Mezzanine where occupants typically 
traversed to the escalator in order to descend to the Concourse; (b) Stairwell B only required one 
(relatively short) horizontal transfer section, while Stairwells A and C required multiple, sometimes 
lengthy (over 1 00 ft) horizontal transfers; (c) emergency response personnel preferentially used Stairwell 
B to climb to higher floors; and (d) an occupant may have switched stairwells during egress, introducing a 
significant uncertainty. Therefore, these four factors confound any conclusions regarding stairwell width 
which may be drawn from the evacuation of the WTC towers. Respondents reported three pieces of 
information critical to this analysis: number of floors they had to climb down, the total time spent in the 
stairwells, and which stairwell they used. Each reported time was normalized (divided by) with the 
number of floors descended in order to compare results independently of starting location. The 
normalized times were then averaged over all occupants who reported using that stairwell. 

In Stairwell B in WTC 1, the average occupant spent approximately 61 ± 38 s per floor, while in 
Stairwells A and C (the narrower stairwells), the average occupant spent 53 ± 34 s per floor. The 
uncertainty was calculated using a standard deviation. Unfortunately, the large uncertainty, relative to the 
average, in the data collected from WTC 1, combined with the confounding conditions referred to 
previously, do not allow differentiation of stairwell movement speeds for the wider stairwell compared to 
the narrower stairwells. 

10.2 FACTORS INFLUENCING TOTAL EVACUATION TIME 

The telephone interview results permit rigorous statistical analysis of factors that played a role in the 
evacuation process. In order to understand what factors influenced the overall evacuation time of the 
average occupant in WTC 1 or WTC 2, two primary dependent variables were predicted: how long an 
individual delayed initiating his or her evacuation, where initiation is defined as entering a stairwell or 
elevator with the intention of exiting the building; and how long an individual spent traversing the 
stairwells. Elevator travel time was not considered in this analysis. The sum of these two elements of the 
evacuation process is the total evacuation time. Multivariate regression modeling was utilized to establish 
factors which contributed to overall evacuation time. A more detailed discussion of both causal models, 
including methods, equations, and significance testing can be found in Appendix C: Causal Modeling. 
This chapter will only present the basic model structure and findings. 

1 30 NIST NCSTAR 1-7, WTC Investigation 



Discussion and Analysis 



10.2.1 Predicting Evacuation Delay in World Trade Center 

The analysis of evacuation behavior consisted of two stages. The first stage, which focused on how long 
it took for people to begin evacuating, determined the factors (variables) and social processes (the major 
paths of causal links between variables) that influenced people delaying the initiation of their evacuation 
out of WTC 1 and WTC 2 on September 11, 2001. Evacuation delay was defined as the number of 
minutes that passed from when a person first became aware that something was wrong until he or she 
began evacuating. 

Model Description 

The model used to predict important factors in evacuation delay in the towers used variables that 
preliminary analyses suggested were salient and closely followed general evacuation theory from the 
social and fire sciences The model is diagrammatically illustrated in Figure 10-5. It can be described as 
follows: (1) delay in evacuation initiation (X 7 , the dependent variable in this analysis) was a direct 
consequence of environmental cues (Xi), the floor on which occupants were located (X 2 ), obtaining 
information without seeking it (X 3 ), perceived risk (X 4 ), seeking additional information (X 5 ), and taking 
pre-evacuation actions (X 6 ); (2) taking pre-evacuation actions was a direct consequence of environmental 
cues, floor, obtaining information without seeking it, perceived risk, and seeking additional information; 
(3) seeking additional information was a direct consequence of environmental cues, floor, obtaining 
information that was not sought, and perceived risk; and, finally, (4) perceived risk was a direct 
consequence of environmental cues, floor, and obtaining information without seeking it. This model was 
parsimonious (in other words, the model was constructed of no more or fewer components than were 
necessary), and consistent with theory that stems from research on existing evacuation and risk 
communication research, and the model accurately reflected findings from preliminary analyses on the 
many variables that could have impacted evacuation delay. With this conceptual model, a set of equations 
for evacuation delay, pre-evacuation actions, seeking additional information, and perceived risk that were 
assumed to depend linearly on the factors above were solved simultaneously to obtain the relative 
importance of each factor in predicting the four primary variables. 

Results 

The estimated parameters of the models for WTC 1 and WTC 2 revealed that the model had a very high 
degree of success in explaining evacuation initiation delay, pre-evacuation actions, seeking information, 
and perceived risk in both towers. The adjusted explained variance (R 2 ) for perceived risk was 55 and 
60 percent in WTC 1 and WTC 2, respectively. The adjusted explained variance for sought information 
was 25 percent in both towers. Respectively, the adjusted R 2 for pre-evacuation actions was 68 and 
69 percent for WTC 1 and WTC 2. Finally, the adjusted R 2 for delay in evacuation for WTC 1 and 
WTC 2 was, respectively, 49 and 56 percent. These were extraordinarily high levels of adjusted explained 
variance to observe in a study of human evacuation. The combined amount of variance in the dependent 
variable, evacuation thereby, established the strong predictive power of the models for both towers. All of 
the equations in the model for WTC 1 and WTC 2 were statistically significant at the 0.001 level or 
higher, indicating a high degree of confidence in the explanatory power of the model. 



NIST NCSTAR 1-7, WTC Investigation 1 3 1 



Chapter 10 



/ 



/ 



Perceived 
Risk 



Pre-Evacution 

Actions 

fc.) 



Environmental 
Cues 




/ 



e 7 



Delay 

Initiating 

Evacuation 

(x,) 



Obtained 

Information 

M 



Sought 
Information 



Figure 10-5. Model of evacuation delay. 



Direct Effects 



With a few exceptions, similar findings regarding evacuation delay emerged in both WTC 1 and WTC 2. 
Repeatability of findings lends validity to the conclusions that can be drawn from the analyses. The 
findings are presented below. First, each of the four equations in both models for both towers was 
considered; next, the models were interpreted as a whole so that the most significant paths of influence for 
each tower could be distinguished. 

Predicting Perceived Risk 

The findings that emerged regarding predicting the risk that people perceived were virtually identical 
across the two towers. 'Perceived' risk was considered more important than objective or actual risk 
because identical observations may be interpreted differently by individuals, and because people act on 
the basis of perceptions, or what they believe to be true. The R 2 for perceived risk was 55 percent in 
WTC 1 and 60 percent in WTC 2. In WTC 1, both environmental cues and floor had strong and similar 
impacts on predicting perceived risk, while obtained information had a weaker but statistically significant 



132 



NIST NCSTAR 1-7, WTC Investigation 



Discussion and Analysis 



impact. In WTC 2, once again, both environmental cues and floor had strong impacts on predicting 
perceived risk, and, once again, obtained information had a weaker but statistically significant impact. 

These findings suggest that the risk that people perceived before they began their evacuation increased 
largely as a function of their starting floor (distance to safety) and being exposed to environmental cues. 
Clearly, information had a lesser impact on risk perception than did the two more salient variables of 
experiencing environmental cues and floor height. The only difference in findings between the towers 
was that, in WTC 2, floor height was by far the strongest predictor of perceived risk, while in WTC 1, 
both floor and environmental cues were equal predictors. In other words, those who were on higher 
floors felt more at risk and moved more rapidly to evacuate. 62 

Predicting Seeking Information 

The variable 'sought information' may include verification of the nature of the attack, instructions or 
directions, or other information deemed relevant prior to evacuation initiation. Once again, the findings 
that emerged for predicting seeking information were almost identical across the towers. Explained 
variance (R 2 ) for seeking information was 25 percent in both WTC 1 and WTC 2. In WTC 1, 
environmental cues and floor both had the strongest and identical impacts on seeking information; 
obtained information had no statistically significant impact on seeking information; and perceived risk 
had a slight impact on seeking information. In WTC 2, environmental cues had the strongest impact on 
seeking information; the impacts of floor and obtained information were not significant; and perceived 
risk had a slight impact on seeking information. 

Seeking information in times of rapid onset emergencies is a typical human response, since people need 
to interpret and make sense out of an event before they act on it. The finding that the variable of 
environmental cues was the strongest predictor of seeking additional information is consistent with this 
theoretical finding about "milling" from past research. Obtained information had no impact on seeking 
information in either tower. This was likely because information to make sense out of the event had 
already been obtained. Perceived risk had a similar effect on seeking information — albeit lesser of an 
effect than environmental cues — in both towers. Interestingly, floor height (or distance from the exit) had 
a significant effect on seeking information in WTC 1, but not in WTC 2. Evacuation theory would 
predict that this effect would be present; it was present in the tower that was struck first, and it was not 
present in the tower struck second. 

Predicting Pre-Evacuation Actions 

Pre-evacuation actions included activities such as making telephone calls, gathering belongings, or 
performing any other tasks deemed necessary prior to evacuation initiation. The explained variance (R 2 ) 
for taking pre-evacuation actions in WTC 1 was 68 percent, and it was 69 percent in WTC 2. In both 
towers, the strongest predictor of taking pre-evacuation actions was floor. Environmental cues was also 
predictive of pre-evacuation actions. Once again, observing cues indicating that one is at risk and being 
high in the building with a longer path to safety emerged as strong predictors. Obtaining information had 
virtually no impact on pre-evacuation actions in either tower. Seeking information impacted pre- 
evacuation action likely because the information obtained supported the need to evacuate and, hence, 



Recall, however, that individuals who were not motivated to initiate evacuation on higher floors in WTC 2 were more likely to 
be trapped above the aircraft impact region when WTC 2 was attacked. 

NIST NCSTAR 1-7, WTC Investigation 1 33 



Chapter 10 



related to getting ready to leave. Finally, the impact of perceived risk on taking pre-evacuation actions 
was higher in WTC 2 than in WTC 1 . 

Predicting Delay in Evacuation 

Explained variance (R 2 ) in evacuation delay was 49 percent in WTC 1 and 56 percent in WTC 2. The 
greatest predictor of evacuation delay in both towers was taking pre-evacuation actions. Obviously, doing 
anything before initiating evacuation — including preparation to leave — delayed departure. Setting this 
factor aside, some clear differences emerged between the two towers in terms of the relative impacts of 
the remaining variables in the model. Perceived risk had no direct effect on evacuation initiation delay. 
This finding is consistent with general evacuation theory where perceived risk's impact on actual 
behavior is indirect through other factors. The three factors with the strongest direct effects on evacuation 
delay were the same in both towers. These were environmental cues, floor, and obtained information. In 
both towers, floor's effect was negative, that is, the more floors one was from the exit, the quicker people 
were to initiate their evacuation. Environmental cues and information that was received passively both 
increased delay in the initiation of evacuation. Finally, seeking additional information had a minimal 
impact on evacuation delay. 

Paths of Greatest Influence and Conclusions 

Although each of the aforementioned findings are interesting in and of themselves, perhaps the most 
important findings that we can offer are those that emerge when all of the individual findings offered 
above are brought together and viewed at the same time in the context of the entire model. 

Bias 

As discussed previously, any conclusions about evacuation initiation delay time in WTC 2 should 
consider the impact of disproportional decedent location, particularly as a source of the disproportionality 
may be highly correlated to the variable of interest, evacuation delay. In other words, those who 
exhibited long delay times in one region of the building were unable to be interviewed, thus artificially 
shortening the average delay time for one-third of the building. In the causal modeling, this would affect 
the relationship between 'floor' and 'delay initiating evacuation,' likely tending towards zero a slightly 
negative estimate (-0.19) of the beta value between the two variables. As floor was not a primary path 
which directly predicted evacuation initiation delay in WTC 2, the impact of this bias was considered 
secondary. The effect of this bias as it worked through other variables was not considered. 

WTC1 

Although there were other factors that had some lesser impacts on influencing what people did, the paths 
of causal influence that defined the main process that led to delay in the evacuation of WTC 1 on 
September 11, 2001 follows. 

Environmental cues (information from the physical environment that something was terribly wrong) and 
floor (increased distance to safety) caused people to set out to find additional information, most likely 
information about what was going on and what they should do. Next, the act of seeking additional 
information, that is, "milling" about to make sense out of the situation, led people to take actions to 



1 34 NIST NCSTAR 1-7, WTC Investigation 



Discussion and Analysis 



prepare to evacuate. Finally, taking those actions to prepare to evacuate delayed the initiation of actually 
evacuating. 

In addition to this four step causal process, environmental cues and floor also had indirect impacts on 
evacuation delay as follows. Both factors increased the odds of seeking information and both factors 
increased the chances that people would take pre- evacuation actions prior to evacuating. Both factors also 
had direct impacts on actual evacuation delay. Environmental cues increased delay while floor decreased 
delay. 

WTC 2 

Although there were other factors that had lesser influence on what people did, the paths of causal 
influence that defined the main process that led people to delay in the evacuation of WTC 2 on 
September 11, 2001, were identical to WTC 1 with one decided difference. 

Environmental cues (information from the physical environment that something was terribly wrong) and 
floor (increased distance to safety) predicted perceived risk. Environmental cues, floor, and perceived 
risk caused people to set out to find additional information. Next, the act of seeking additional 
information, that is "milling" about to make sense out of the situation, and perceived risk both led people 
to take actions to prepare to evacuate. Finally, taking those actions to prepare to evacuate delayed the 
initiation of actually evacuating. 

In addition to this five step causal process, environmental cues and floor also had indirect impacts on 
evacuation delay. Both factors increased the odds of seeking information and both factors increased the 
chances that people would take pre-evacuation actions prior to evacuating. Both factors also had direct 
impacts on actual evacuation delay. Environmental cues increased delay, while floor decreased delay. 

10.2.2 Predicting Normalized Stairwell Evacuation Time in WTC 1 on 
September 11, 2001 

The second component of total evacuation time was the time spent in the stairwells. This analysis 
determined the factors and social processes that influenced the normalized stairwell evacuation time per 
story of stairs for the people who evacuated out of WTC 1 on September 11, 2001. WTC 2 was excluded 
from this analysis because evacuees used stairs, elevators, and/or a combination of both for their 
evacuation and could not be separated for the analysis. Evacuation time was defined as the average 
number of seconds per story of stairs that it took people from the time they entered a stairwell until they 
completed their evacuation out of the building. The model used to predict important factors in stairwell 
evacuation time again used variables that preliminary analyses and general evacuation theory suggested 
as salient, and is illustrated in Figure 10-6. It can be described as (1) normalized stairwell evacuation time 
is a direct consequence of floor, evacuation decision delay, environmental cues, emergency responders, 
crowding, and evacuation interruption; (2) evacuation interruption is a direct consequence of floor, 
evacuation decision delay, environmental cues, emergency responders, and crowding; (3) crowding is a 
direct consequence of floor, evacuation decision delay, environmental cues, and emergency responders; 



NIST NCSTAR 1-7, WTC Investigation 1 35 



Chapter 10 



Evacuation 

Decision 
Delay 



Emergency 

Responders 



Evacuation 
Interruption 




Normal 

Stairwell 

Evacuation 

Time 



Environmental 

Cues 

x, 



Crowding 




Figure 10-6. Causal model for predicting normalized stairwell evacuation time. 

(4) emergency responders is a direct consequence of floor, evacuation decision delay, and environmental 
cues; (5) environmental cues is a direct consequence of floor and evacuation decision delay; and, finally, 
(6) evacuation decision delay is a direct consequence of floor. This model is parsimonious, consistent 
with the high rise fire evacuations theory, and it well-represented the positive findings of our preliminary 
analyses of the many variables that could have impacted normalized stairwell evacuation time. 

Results 

The estimated parameters of the model for WTC 1 revealed that the model had a very high degree of 
success in explaining normalized stairwell evacuation time. The adjusted explained variance (R 2 ) for 
normalized stairwell evacuation time was 44 percent, 1 1 percent for evacuation interruption, 72 percent 
for crowding, 57 percent for emergency responders, 79 percent for environmental cues, and 34 percent for 
evacuation initiation delay. With the exception of evacuation interruption, these are extraordinarily high 
levels of adjusted explained variance to observe in a study of human evacuation; and, these R 2 's, thereby, 
establish the strong predictive power of the model. All of the equations in the model were statistically 
significant at the 0.001 level or better. 

Direct Effects in the Model 

We first consider each of the factors individually, and then the model is interpreted as a whole so that the 
most significant paths of influence can be distinguished. 



136 



NIST NCSTAR 1-7, WTC Investigation 



Discussion and Analysis 



Predicting Evacuation Initiation Delay 

Evacuation initiation delay is the total time from first awareness that something was wrong on 
September 11, 2001, until entering a stairwell to evacuate. The findings that emerged regarding 
predicting delay in the initiation of evacuation from floor were that the R 2 was 34 percent. This 
relationship is already discussed in detail in the previous section. 

Predicting Environmental Cues 

Environmental cues were the visual, auditory, or other sensory perceptions that indicated danger on 
September 11, 2001. Explained variance (R 2 ) for observing environmental cues was 79 percent, and the 
equation was statistically significant at the 0.001 level. Floor had a very strong direct impact on observing 
environmental cues, and it was significant at the .001 level. The effect of delay in the initiation of 
evacuation was smaller, but still statistically significant at the 0.001 level. It would appear that the longer 
a person took to begin their evacuation, the more the physical impacts of the event grew and the more 
likely people were to experience them. 

Predicting Emergency Responders 

Emergency responder was denoted as either observed or not observed. The explained variance for 
predicting encountering emergency responders was 57 percent, and the equation was statistically 
significant at the .001 level. Experiencing environmental cues and floor both predicted encountering 
emergency responders. This makes sense when one considers that emergency responders would be most 
likely to go to areas experiencing the impacts that would also yield environmental cues, and the higher 
one was in the tower, the more stairwells one had to traverse, the longer the traversal time overall, and, 
thus, the greater the odds of encountering emergency responders. 

Predicting Crowding 

Stairwell crowding was denoted when a respondent indicated that the stairwell was crowded to the extent 
that it impacted progress down the stairs. The explained variance for perceived crowding on the 
evacuation stairwells was 72 percent. Perceived crowding largely increased as a result of environmental 
cues and encountering emergency responders. 

Predicting Evacuation Interruption 

Evacuation interruption was denoted by respondents who indicated that they left the stairwell or stopped 
in the stairwell by choice. The explained variance for interrupting evacuation was only 1 1 percent. Even 
though this equation was statically significant at the .001 level, none of the five predictor variables in the 
equation had a statistically significant impact on evacuation interruption. At best, these findings can be 
interpreted to mean that there was a slight but not statically significant tendency for people to interrupt 
their evacuation if they had more rather than fewer floors to traverse to safety, and if they encountered 
environmental cues (perhaps obstacles) in the process of evacuation. 



NIST NCSTAR 1-7, WTC Investigation 1 37 



Chapter 10 



Predicting Normalized Stairwell Evacuation Time 

Explained variance in predicting normalized stairwell evacuation time was 44 percent, and the equation 
was statistically significant at the .001 level. Only two factors in the equation were statistically 
significant, both at the .001 level. These were environmental cues and evacuation interruption. Clearly, 
the single factor that had the biggest impact on increasing the amount of time people spent, on average, 
per stairwell was environmental cues. The only other factor that had a significant impact was interrupting 
evacuation, obviously, because stopping egress would increase the amount of time needed to complete 
evacuation. 

Paths of Greatest Influence and Conclusions 

Although each of the aforementioned findings are interesting in and of their own right, the most important 
findings that we can offer are those that emerge when all of the above findings are brought together and 
viewed at the same time in the context of the model as a whole. Thus, the main process that led to 
increased normalized stairwell evacuation time in the evacuation of WTC 1 on September 1 1th was 
straightforward and clear. Floor (increased distance to safety) substantially increased the odds that people 
would encounter environmental cues. Floor also increased delay in starting evacuation (this relationship is 
elaborated upon in much detail previously), which, in turn, also increased the chances that people would 
encounter environmental cues. But it was encountering environmental cues that had a large and direct 
effect on increasing the amount of time that people spent, on average, to traverse their evacuation 
stairwell. In addition to this multi-step process with environmental cues as the key predicting variable, 
interrupting the process of evacuation for any reason also increased the amount of time, on average, that 
people used to descend their evacuation stairwell. 

10.3 ISSUES THAT IMPACTED OCCUPANT EVACUATION 

While the multivariate regression analysis of the telephone data has extraordinary power to distill the 
telephone interview data to the salient features, the analysis was necessarily limited by the constraints of 
the number and type of questions contained in the telephone interviews. Face-to-face interviews and 
focus groups provided a more detailed understanding than can be achieved by exclusive use of the 
telephone interviews. This section provides a discussion of some of the issues identified from both the 
telephone survey and face-to-face interviews that impacted the ability of occupants to successfully 
evacuate the two towers on September 1 1, 2001, including both those issues that aided evacuation and 
those that made evacuation more difficult. 

10.3.1 Environmental Cues and Information 

In any emergency, occupants obtain information about the event by passively receiving it and actively by 
seeking it out through various means. Information may have led occupants to evacuate or to delay 
evacuation to gather additional information or to perform additional tasks prior to beginning evacuation. 
As soon as WTC 1 was struck, most occupants' first reaction was to look out the window for information. 
These first environmental cues from the building (hearing the explosion, feeling the building shake, etc), 
coupled with the reaction of looking out the window (and seeing the fireball and smoke in WTC 1 , the 
debris falling outside of the window, or even the plane hitting the tower), was enough to signal that 



138 NIST NCSTAR 1-7, WTC Investigation 



Discussion and Analysis 



something was wrong for many in the towers. This first set of information caused some to begin the 
evacuation process immediately. 

An occupant in WTC 1 from a floor in the 70s reported immediate 
evacuation initiation: "The building was swaying more than normal and 
I slammed into the cabinets. I got up and headed for the staircase." 
Interview 1000574 (NIST 2004) 

Also, an occupant in WTC 2 from a floor in the 80s also noted 
evacuating immediately. "Somebody in my area jumped up and said 
there is a bomb. At this stage, everybody else took the stairs. I went 
directly to the Stairwell C. I didn't think I was in danger at this point." 
Interview 1000568 (NIST 2004) 

In both towers, many delayed evacuation to find out more information before entering a stairwell or 
elevator. This may have involved talking with coworkers, making and receiving phone calls, sending and 
receiving emails, searching the internet, and watching TV. 

According to the telephone interviews, 1 1 percent of the occupants in WTC 1 and 21 percent of the 
occupants in WTC 2 received information related to the event through various means. Of the occupants 
who received information, 57 percent in WTC 1 and 65 percent in WTC 2 received information about 
what happened. A majority of the occupants in WTC 2 (54 percent) received information about what 
happened to WTC 1 through the public address announcement. Many of them, already in the process of 
evacuating, were faced with the decision of whether to return to their offices or continue their evacuation. 

While on their floor, occupants were receiving information about what was going on in the buildings from 
each other and through a variety of technological means. Technology such as landline phones, cell 
phones, blackberries, email, internet, and TV was used inside the building to receive such information. 

An occupant from a floor in the 80s of WTC 1 received a phone call 
from a co-worker. "A co-worker was calling from home. She said that 
'a plane hit the building and you better get out of there.'" Interview 
8000005 (NIST 2004) This information caused the occupant to begin 
their evacuation process. 

However, the situations in the towers were different before WTC 2 was hit. The information that was 
received by the occupants in both buildings was that only WTC 1 had been hit by an aircraft. This 
message produced a variety of responses from occupants in WTC 2. 

After hearing the public address system announcement, an occupant from 
a floor in the 50s in WTC 2 was calling an out-of-town office to let them 
know they were evacuating the building. "I got a hold of them and they 
told me that a plane hit the building [WTC 1] and to get out of there." 
The respondent hung up and continued to call loved ones and answer 
other phone calls until his or her own building was hit by an aircraft. 
Interview 1001666 (NIST 2004) 

An occupant in the 40s from WTC 2 was watching TV before WTC 2 
was hit. "I saw what happened on CNN and was informed that a plane 



NIST NCSTAR 1-7, WTC Investigation 1 39 



Chapter 10 



hit Tower 1. I entered the emergency stairwell A and proceeded down. 
Interview 1000867 (NIST 2004) 

Even after entering the stairs, the evacuation flow continued through the use of technology and 
information sharing among the occupants inside the building. Although they had already entered the 
stairs, additional information (aside from the public address system announcement in WTC 2 telling 
occupants to return to their floor) sometimes increased the occupants' motivation to evacuate the 
building, as shown in the following quotes from WTC 1 occupants. 

An occupant from a floor in the 50s in WTC 1 was traveling in a 
stairwell when he or she heard another occupant say "that another plane 
hit the building. [The person] got this information from a cell phone and 
we speeded up our evacuation." This respondent went on to indicate that 
"cell phones were working and people were trying to figure out what was 
going on." Interview 1000071 (NIST 2004) 

While in Stairwell B, an occupant from a floor in the 70s in WTC 1 
"heard a fire department radio say the words '. . . Structural Instability . . 
. '" This radio message was relayed at the time of the collapse of 
WTC 2. The respondent went on to say that they "stopped because there 
was no movement in the staircase, but then became frightened for the 
first time. [I] began to focus more on getting out." Interview 1000118 
(NIST 2004) 

An occupant in WTC 2 also experienced the same knowledge about his 
building during evacuation. "The phone was ringing, [so] I answered the 
phone on the 3rd floor. I listened to my daughter tell me that both 
buildings were hit [by a] commercial airliner, and that it was a terrorist 
attack. I told my boss what I learned from my daughter, so we would 
both have more determination to get out." Interview 1000003 
(NIST 2004) 

When people were given new information about the event from outside of the building, they shared that 
information freely with those around them. People provided information about the phenomena that they 
witnessed themselves (at first awareness) or information that they had received from another source, such 
as from someone outside of the building. As evidenced in face-to-face interviews, the behavior of sharing 
information was frequent, regardless of whether the individuals involved were strangers. 

For example, there was an occupant in WTC 1, who originated on a floor 
in the 40s, whose phone rang while in the stairwell: "I answered the 
phone and my [spouse] was telling me that another plane had struck 
Tower 2. [The spouse] said to get out of the building as fast as you can 
and that it is a terrorist attack. [The spouse] also told me about the plane 
hitting the Pentagon. [The spouse] mentioned that there was another 
plane in the air but [the spouse] wasn't sure where it was. I told others 
around me what had happened, in order to spread the available 
information." Interview 1000572 (NIST 2004) 

In addition to passively receiving information from inside the buildings, 28 percent of the survivors in 
WTC 1 and 29 percent of the survivors in WTC 2 actively sought out information related to the event 

1 40 NIST NCSTAR 1 - 7, WTC Investigation 



Discussion and Analysis 



through various means. In the face-to-face data, this includes nearly everyone when the occupants who 
looked out the window for information are included. However, the telephone interview data shows that 
approximately 40 percent of these who sought out information about the event in both towers were not 
successful in finding the information they were searching for. 

For example, an occupant from a floor in the 60s of WTC 1 tried making 
several phone calls after the building was hit, but was unsuccessful. 
"The building started to sway and everything started shaking. I knew 
that there was something [wrong]. I ran to my desk and made a couple 
of phone calls. I dialed about five times trying to reach my [spouse]. I 
also called my sisters to find out more information." Interview 1000733 
(NIST 2004) 

Occupants also attempted to seek information about what was going on in the buildings through a variety 
of technological means. This technology included landline phones, cell phones, blackberries, email, 
internet, and television broadcasts inside the building. 

An occupant in the 70s performed a variety of information seeking 
activities before deciding to leave WTC 1 . "I was curious and wanted to 
get information on what happened. I tried to find information on the 
internet, [because] that was the fastest way of getting information. 
[Also], I knew my mother had TV access. I called my mother to get 
information on what she had seen so far on TV." This occupant also 
phoned a brother-in-law for more information until the occupant 
witnessed the plane impact WTC 2. Then, the occupant began to 
evacuate. Interview 1000583 (NIST 2004) 

Also, an occupant from a floor in the 100s in WTC 2 noted that "there 
were a lot of people trying to find out what was going on by using cell 
phones." Interview 1000563 (NIST 2004) 

Emergency responders also confirmed what happened to inquiring occupants inside the stairwell, but then 
reassured them that everything would be okay. Obtaining information from emergency officials inside 
the building was not frequently reported. In return, occupants would provide information to the 
firefighters about where certain people were waiting for rescue assistance. 

Overall, the information flow and its impact on occupants' evacuation were dependent upon many factors, 
including the tower in which the occupant was located and the time the information was received during 
the emergency. Occupants in WTC 2 were sometimes affected differently to news that only WTC 1 had 
been hit by aircraft, when compared to occupants in WTC 1. Information was received and sought out at 
various times throughout an occupants' evacuation and from different sources. However, many 
occupants in both towers reacted to the first set of information provided by the plane crash into WTC 1, 
either by initiating evacuation or by beginning pre-evacuation activities. 

10.3.2 Building Alarm Systems 

Fire alarms have been a hallmark of fire protection for more than a century, with systems proposed in 
New York City as early as 1847 (Bukowski and Moore 2003). Typically, the systems serve dual purposes, 



NIST NCSTAR 1-7, WTC Investigation 141 



Chapter 10 



occupant notification and transmission of information about the fire to emergency responders. On 
September 11, 2001, with significant damage to the fire alarm system in the towers, only a small 
percentage of occupants in both towers (14 percent in WTC 1 and 10 percent in WTC 2) noted hearing a 
fire alarm during different stages of their evacuation. Many occupants in WTC 1 had already begun to 
evacuate their building due to a variety of building observations from the plane impact (less than 
5 percent of occupants in WTC 1 said that their decision to evacuate was based on the fire alarm going 
off). The same is true for occupants in WTC 2 who heard alarms after their tower was hit, since a large 
majority had already begun evacuation before or when their tower was hit (less than 1 percent said their 
decision to evacuate was based on hearing a fire alarm). 

Some occupants of WTC 2 believed that the emergency was limited to WTC 1 and were unclear if their 
evacuation was necessary. Also, occupants in WTC 2 were faced with a public address system 
announcement informing them to return to their offices. For example, a respondent in WTC 2 heard the 
announcement that the building was secure first, followed by the fire alarm. 

"At the 25th floor, I heard someone on the loudspeaker who said that 1 
World Trade was hit but that 2 World Trade was OK and that we should 
all go back to our office. After I heard that announcement, I heard the 
general alarm which meant to me to leave the building even quicker." 
Interview 1000740 (NIST 2004) 

And, in response to this, the occupant kept walking down the stairs to leave the building. 

In both towers, occupants reported hearing fire alarms at various floors throughout the tower. For those 
occupants in WTC 1 and WTC 2 who heard fire alarms on their floors, the fire alarm was not frequently 
nor specifically mentioned as the reason for evacuating. Instead, the top three reasons for occupants to 
begin evacuation from either towers were the plane hitting WTC 1 , being told to evacuate, or feeling 
afraid or in danger (in no particular order for each tower). 

10.3.3 Public Address Announcements 

Public Address announcements were not mentioned in the face-to-face interviews as being heard after 
each building was hit by aircraft. However, from the telephone survey, 14 percent of the survivors in 
WTC 1 reported receiving information from a public address system announcement - which suggests that 
the occupants received an announcement after their tower was struck. There is also information from the 
9-1-1 tapes that a second announcement was heard in WTC 2 by people near and above the floors of 
impact after the building was hit, although NIST found no evidence from face-to-face interviews that 
survivors below the floors of impact heard this announcement. 

A building-wide public address announcement was made to the occupants in WTC 2 approximately three 
minutes before the aircraft hit their building. Based upon analysis of many face-to-face interviews and 
published accounts, the first announcement provided to occupants in WTC 2, before their tower was hit, 
stated the following: 

There is a fire condition in WTC 1 , WTC 2 is secure/safe, please return 
to your offices. 



142 NIST NCSTAR 1-7, WTC Investigation 



Discussion and Analysis 



Fifty-four percent of the occupants in WTC 2 reported receiving information from the public address 
system. Since 40 percent of WTC 2 survivors had evacuated before WTC 2 was attacked, the 
overwhelming majority of occupants still in the building heard the 9:00 a.m. announcement. While some 
occupants interpreted the first announcement as a suggested course of action, other interpreted it as an 
instruction. 

Approximately two minutes later, at 9:02 a.m., one minute before WTC 2 was attacked, a second 
announcement was made, contradicting the first announcement. Recorded in the background of an 
answering machine receiving a call from an occupant calling home from floor 98, the announcement 
indicated that it was now permissible to initiate an evacuation: 

"May I have your attention please. The situation is in Building 1. 
However, if conditions on your floor warrant, you may wish to start an 
orderly evacuation." Interview 3000001 (NIST 2004) 

The 9:02 a.m. announcement was noted by one occupant catalogued in the collection of published 
accounts (Fahy and Proulx 2003), however, the large discrepancy in the percentage of occupants 
reporting this announcement when compared to the percentage who reported the 9:00 a.m. announcement 
remains unresolved. 

The majority of occupants in WTC 2 began their evacuation before the first announcement was made. 
Some of them never heard the announcement (due to evacuating via elevator, starting in the lobby, or 
possibly evacuating from a lower floor before the announcement was given) and some heard the 
announcement from either the stairs or a lower floor than their original starting floor. It is known that 
occupants listened to the announcement from the stairs, left the stairs to hear the announcement more 
clearly, or were led out of the stairs minutes before the announcement on skylobby floors by security 
personnel. Those who heard the announcement and had begun the evacuation process were faced with a 
decision on whether to comply with the instruction and return to their offices or continue their evacuation. 
The face-to-face interview data shows that while many people did not follow the instruction/suggestion in 
the announcement, others chose to return to their office. Each individual's decision was affected by a 
number of factors at the time of the announcement, including: 

• their perception of danger and ability to evacuate: 

"While I was on the steps, I heard a public address system announcement 
that said the building was secure and to return to our offices. It made me 
angry to hear this announcement because I felt we were in danger." 
Interview 1000048 (NIST 2004) This occupant, who had a medical 
disability, was helped down the stairs by a fellow occupant, until they 
took an elevator to the ground floor. 

• the actions or statements of others in the immediate area (group behavior): 

An occupant from a floor in the 70s, who used both an elevator and a 
stairwell, heard an announcement around floor 20: "[I heard] a PA 
(public address) announcement that told people to return to their offices. 
I shouted to people to continue down the steps; to keep the flow of traffic 
going—so that people wouldn't be going past me. I intentionally lied to 



NIST NCSTAR 1-7, WTC Investigation 143 



Chapter 10 



people. I heard the PA [public address] announcement tell people to 
return to their offices, but I thought it was better for them to evacuate the 
building and I knew that if the people didn't turn around and try to move 
past me that I could evacuate faster." Interview 1000024 (NIST 2004) 

An occupant from a floor in the 60s in WTC 2: "I heard a public address 
system announcement that said to stay where we were—that our building 
was secure and the problem was with the other building, but the fire 
warden said to ignore the message and to leave the building anyway." 
Interview 1000050 (NIST 2004) 

An occupant from a floor in the 90s heard an announcement around floor 
78 in WTC 2: "We heard an announcement that said 'the building is safe; 
you can go back.' I continued walking down the stairs, to keep moving, 
to not hold up the others that were behind me." Interview 1000070 
(NIST 2004) 

• desire to follow instructions: 

An occupant from a floor in the 30s in WTC 2: "The PA (public address) 
announcement informed us, '. . . You may return to your office.' I exited 
the stairwell on the [a floor in the 30s] and proceeded to the elevator 
lobby on that floor. I entered the elevator and took it to [a floor in the 
30s] to return to my office as instructed. I [then] called my [spouse] at 
about 9:03 a.m. to tell [the spouse] that everything was ok - but my 
conversation was cut short when the second aircraft struck Tower 2." 
Interview 1000049 (NIST 2004) 

• and even serendipity: 

An occupant from a floor in the 10s in WTC 2: "I heard an 
announcement saying that the problem is in Building One and for us to 
stay where we are or to return to our floors. I stepped off of the stair into 
the 9th floor to take the elevator to return to my floor. I was distracted 
because of conversation and did not realize that the elevator was going 
down until it started moving." Interview 1000922 (NIST 2004) This 
occupant did not return to their office, but instead, evacuated the building 
after exiting the elevator. 

Occupants who were led out onto skylobby floors also hesitated after hearing the announcement. They 
engaged in milling with other coworkers in an attempt to decide what to do. Minutes later, WTC 2 was 
struck, which caused them to initiate their evacuation. 

Other occupants who did not begin the evacuation process were still located on their floor when the PA 
(public address) announcement was made. The announcement helped to reinforce their original goal of 
remaining on their floor for some time. Most of these occupants listened to the announcement and 
remained on their floor. 

"I heard an announcement on the intercom that told us to stay. I decided 
to stay to avoid getting in more danger. I heard the phones ringing off 



144 NIST NCSTAR 1-7, WTC Investigation 



Discussion and Analysis 



the hook and my boss and I were going from phone to phone answering 
them. This was to reassure the families that everyone was ok." 
Interview 1000524 (NIST 2004) 

Regardless of where occupants were located in the building and whether they had decided to begin the 
evacuation process, their final evacuation decision was prompted by the impact of the aircraft into their 
building, WTC 2, at 9:02:59 a.m. 

10.3.4 9-1-1 System 

NIST was given access to all emergency calls to the New York City 9-1-1 system related to the World 
Trade Center attacks on September 11, 2001 from 8:45 a.m. until shortly after 10:30 a.m. The 9-1-1 
system was flooded with calls from 8:46:30 a.m., when WTC 1 was attacked, until roughly 8:55 a.m., and 
experienced a second surge in call volume at 9:03 a.m., when WTC 2 was attacked, which continued for 
several minutes. The majority of calls were eyewitnesses reporting that they had observed something 
crashing into the World Trade Center or reporting that the World Trade Center was on fire. Initial reports 
included descriptions of small planes, helicopters, large (commercial or military) aircraft, as well as a 
bomb or a missile. A few callers sought advice about whether they should evacuate their nearby 
buildings or information about what was going on. Other callers reported observations of suspicious 
activities or people. After several minutes, the first WTC occupants began to call 9-1-1 seeking advice or 
rescue assistance. Occupants continued to use the 9-1-1 system, both as a resource for information and an 
outlet for reporting their situation, until around 10:00 a.m., when WTC 2 collapsed. 

During the time period between aircraft impacts, advice from 9-1-1 operators to occupants experiencing 
smoky conditions was largely to shelter in place and await rescue from emergency personnel. 

One of the common questions asked by trapped occupants of the 9-1-1 operators was whether to break 
windows in order to obtain fresh air. The advice given to occupants varied by operator. Some operators 
encouraged the occupants to assess their own unique situation and determine whether breaking a window 
would help, sometimes warning the occupants that breaking a window may only serve to introduce more 
smoke from the outside than it would relieve from the inside. Other operators simply advised occupants 
not to break windows. The 9-1-1 operators, largely acting without complete knowledge of an evolving 
and traumatic situation, were forced to strike a balance between efficiently logging incoming calls, 
compassionately counseling sometimes desperate occupants, and disseminating relevant event 
information. 

Another common question from occupants to 9-1-1 operators was whether they should evacuate, and 
further, whether that should be upward or downward (in other words, where was the impact region 
relative to their location). Some 9-1-1 operators did not initially offer advice, routinely telling the 
occupants to defer to the instructions they were (presumably) receiving from building authorities. Other 
operators consistently advised occupants to stay in place, rescue was on the way. At least one occupant 
below the floor of impact in WTC 2 was repeatedly advised to await rescue by emergency personnel who 
were advised of the occupant's whereabouts. Sometime after 9:02:59 a.m., some operators began to 
encourage the occupants to evacuate, if possible. Information about the location or magnitude of the 
impact in either WTC 1 or WTC 2 was not generally communicated to trapped occupants. 



NIST NCSTAR 1-7, WTC Investigation 145 



Chapter 10 



Operators routinely collected from callers specific location information, which proved useful during the 
NIST investigation for establishing the time-evolving building conditions (fire, smoke, damage, number 
of people), including: building, floor, corner (northwest corner, for example), or office number. 
Additionally, callers would sometimes relay individual or group observations about the tenability or 
status of the egress path (i.e., access to stairwells or elevators). For these reasons, the 9-1-1 recordings 
proved to be a valuable record of the conditions above the impact floors for which there would have been 
no other method to discover. 

10.3.5 Previous Evacuation Experience 

Using prior evacuation experience to guide future evacuation decisions, may or may not produce better 
outcomes. Recall from Section 4.3 that 16 percent of survivors on September 11, 2001 were also present 
during the 1993 bombing. A survivor from the 70s in WTC 1 described: 

"My response to the '93 bombing wasn't sufficient. I realized there was 
no real purpose in sticking around. I was going to get out as quickly as I 
could." Interview 1000525 (NIST 2004) 

Another survivor from a floor in the 70s in WTC 1, however, used their 1993 evacuation experience to 
justify delaying their evacuation: 

"I was there in 1993, and [this time] I wanted to wait for some directions 
from someone, through the speaker system, fire alarms, etc. I stayed on 
around the general vicinity where I worked. This is the main reason why 
I stayed longer on the floor." Interview 1000576 (NIST 2004) 

Similarly, 16 percent (n=59) of WTC 2 evacuees on September 11, 2001 also evacuated in 1993. In 
WTC 2, however, only 75 percent (n=42) felt that they made the right decision in 1993 (compared to 
95 percent in WTC 1), possibly due to the fact that many more waited to evacuate in 1993 in WTC 2 
(69 percent (n=39)) than did so in WTC 1. Only 31 percent (n=17) who reported their decision evacuated 
immediately from WTC 2 in 1993, keeping in mind that the bomb had a more significant impact on 
WTC 1 in 1993. 

Nearly every respondent who compared the 1993 bombing evacuation to the 2001 evacuation indicated 
that the 1993 evacuation was slower, more difficult (presence of smoke in the stairwells and floors), and 
more stressful. Having participated in the 1 993 evacuation, those occupants generally felt much better 
about the progress and conditions during the 2001 evacuation. As a consequence of their 1993 
experience, however, several 2001 survivors who had direct 1993 experience reported having diminished 
confidence in building announcements (or the lack thereof), exemplified by this occupant from a floor in 
the 40s in WTC 2: 

"I ... was there during the 1 993 bombing. I did evacuate —with a group 
of people who had no clue as to where we were going. What I learned 
from that experience was to not trust the Port Authority's 
announcements. Because of experiencing the '93 bombing, I felt a strong 
conviction to get out this time. Had I not experienced that, I might have 
listened to the Port Authority announcements and stayed put." Interview 
1000048 (NIST 2004) 



146 NIST NCSTAR 1-7, WTC Investigation 



Discussion and Analysis 



10.3.6 Occupant Activities 

Occupants of WTC 1 and WTC 2 performed a number of activities before and during evacuation on 
September 11, 2001. These activities included gathering personal belongings, milling with other 
occupants, seeking additional information, and calling family or friends, among others. In general, these 
activities either delayed the start of their evacuation (pre-evacuation activities) or interrupted their 
evacuation once it was in progress and are described below. 

Pre-Evacuation Activities 

Occupants performed a variety of activities prior to entering the stairwell (or elevator in WTC 2) to begin 
their evacuation. An activity such as gathering occupants or warning them to leave has a variable amount 
of time associated with it, depending upon the floor space to search and how reluctant others are to leave, 
whereas an activity such as gathering belongings usually requires under a minute to complete. Other 
activities noted by occupants were milling, seeking information (such as looking out the window, making 
phone calls, searching the internet), helping others, following emergency procedures, fire fighting, and 
working or closing up a work desk. The delay reported by survivors in starting their evacuation was 
predominately one to five minutes, while a small number delayed for more than an hour. The 
distribution of evacuation initiation delay times was discussed in detail in Chapter 6. 

The majority of the survivors, 70 percent of those in WTC 1 and 75 percent in WTC 2, engaged in the 
activity of talking to others, or milling. The behavior of milling is used by occupants to discuss what they 
witnessed and to arrive at conclusions for what happened and what to do next. 

For instance, an occupant in the 90s of WTC 2 looked for other 
coworkers after he heard a sound coming from WTC 1. "I moved back 
out of my office where I could get in sight with other people to share my 
observations, to discuss what ought to be done, and to hear what the 
consensus would be." Interview 1000689 (NIST 2004) 

After hearing an explosion and feeling the building rock, an occupant in 
the 60s of WTC 1 turned to a coworker. "I discussed with my coworker 
what we were going to do. The noise of the explosion, the shaking of the 
building, and seeing the paper falling was what made me decide it was 
something serious. We left our conference room and went into the 
hallway to get to the stairs." Interview 1000053 (NIST 2004) 

Near the floors of impact in WTC 1 , a few survivors worked to secure an area of refuge or perform fire 
fighting activities. The floors near the impact area were much more difficult to navigate due to the smoke 
and damage. Because of this, some occupants remained in an office, surrounding the openings with wet 
towels, before being rescued by Port Authority personnel. 

In both towers, the face-to-face data captures the fact that some occupants delayed their evacuation start 
to wait to receive evacuation instructions, as part of their emergency procedures. Fourteen out of 
124 respondents in WTC 1 and 12 out of 69 respondents in WTC 2 waited for instructions. The 
percentages are higher for WTC 2 because it included those occupants who heard the public address 
system announcement while still on their floor before they began evacuation. 



NIST NCSTAR 1-7, WTC Investigation 147 



Chapter 10 



On at least one floor in the 30s in WTC 1, occupants followed the 
emergency procedures. "We knew that we were supposed to go out to 
the hallway and that there would be someone there with information or 
information would be given via loudspeaker. We went out to the 
hallway to listen for any instructions." Interview 100009 (NIST 2004) 

Despite the number of pre-evacuation activities of some occupants, most occupants in WTC 2 chose to 
begin evacuating before their tower was attacked. Eighty-six percent to 91 percent 63 of the survivors in 
WTC 2 began their evacuation before their building was hit at 9:02:59 a.m. Similar to the telephone 
interview results (which are generalizable), 12 percent of the face-to-face respondents in WTC 2 stayed 
until their tower was attacked. Of those who stayed (8 interviews out of 66 from WTC 2), five held 
positions of corporate leadership or fire warden. The remaining three made multiple decisions to leave 
before the plane hit, but were delayed by announcements, phones ringing, and the desire to gather 
belongings. 

Helping 

September 11, 2001 showed that people are willing to help out others during an emergency, even if they 
do not know the person ahead of time. Those occupants in need of help during the evacuation included 
occupants who use canes or wheelchairs, overweight, elderly, pregnant women, and people with asthma. 
Many of those who supported occupants in need throughout their evacuation were strangers. These 
helpers often remained with the occupant in need throughout their entire evacuation, even though they 
were putting themselves at risk. Short-term examples of helping behaviors mentioned in both towers 
were comforting coworkers in the stairwell, leading occupants in need to the elevators, helping an 
occupant carry belongings, and encouraging others to keep evacuating in times of stress or exhaustion. 
Occupants also exhibited major feats of heroism on September 11, 2001 by caring for an individual 
throughout their entire evacuation, sometimes assisting strangers down more than 60 stories to the 
building. 

Examples of this heroism include: 

A group of coworkers helped to rescue a wheelchair user from a floor in 
the 60s in WTC 1 . 15 minutes after the impact of their tower, WTC 1 , 
they located the evacuation chair and transferred the occupant into it. 
"We began our descent sliding the evacuation chair, step by step, [with] 
two helpers holding the chair on top and two helpers holding the chair on 
the bottom. [Soon] we began carrying [the occupant] down the 
staircase." Interview 1000123 (NIST 2004) 

Two occupants assisted an overweight occupant in WTC 1 in navigating 
the stairwell from a floor in the 50s. Towards the bottom of the stair, the 
occupant's "legs would give out and [the occupant] would fall in our 
arms. Our descent slowed dramatically as we practically had to carry 
[the occupant] down each flight of stairs." Interview 1000093 (NIST 
2004) 



63 



The discrepancy reflects a small difference among responses to two independent questions: 'How long did it take you to 
evacuate?' and 'Did you evacuate WTC 2 prior to WTC 2 being attacked?', as discussed previously. 

148 NIST NCSTAR 1 - 7, WTC Investigation 



Discussion and Analysis 



On a floor in the 80s of WTC 1 : "I heard the secretary, who was in 
flames, calling for help near the door. I extinguished the flames with my 
bare hands, along with a colleague who was using a sweater." Interview 
1000113 (NIST2004) 

However, an offer of help was not always accepted. One survivor from a floor in the 60s in WTC 1 
recalled: 

"There were some people that stopped in descending because they 
needed to catch their breath. We stopped about three times and offered 
help, which was not accepted." Interview 1000878 (NIST 2004) 

Resting During Evacuation 

Some occupants felt the need to rest at certain points during their evacuation. An occupant coming down 
from the 90s by the stairs was forced to walk roughly one-half mile before eventually leaving the 
building. 64 Occupants often chose to rest inside the stairway, either on the steps or, more frequently, on a 
stairwell landing. Resting during evacuation was noted by respondents beginning evacuation from either 
the high or medium strata of each building. Most often, resting was reported by occupants with 
respiratory problems, obesity, or other physical or medical conditions. 

From the face-to-face interviews, respondents' experience of resting on the stairs is captured in the 
following quotes: 

An occupant in WTC 1 had to travel from a floor in the 70s. Around 
floor 25, the occupant's "legs were really hurting and it was difficult to 
walk. I slowed down and stopped about three times. The first time, I let 
people get ahead of me. I stopped at a landing for approximately two 
minutes. I proceeded down but then stopped for the second time, 
somewhere between the 15th and tenth floors. I waited there for three or 
four minutes to think what was the best way to continue. There was 
water gushing from the fifth floor onto the staircase. I stopped and began 
to think about the safest way to get down and decided to proceed more 
slowly and take a firmer grip on the rail." Interview 1000111 (NIST 
2004) 

In WTC 2, a group of coworkers found themselves also having to rest 
multiple times during their evacuation from a floor in the 90s. "On the 
78th floor, we felt tired. My coworker, who is a diabetic, hadn't eaten 
breakfast yet. We stopped on the stairs to rest for a minute. A lot of 
people seemed to be stopping, probably due to the heat and needing to 
catch their breath. On the 64th floor, another coworker turned to me and 
said [he or she] was getting tired and didn't feel well. I said 'We will 
take a little break' and I gave [the occupant] mints." Interview 1000526 
(NIST 2004) The group stopped at least two additional times because a 



64 



At roughly 30 degree slope, an occupant descending stairs from 1,000 vertical ft also travels over 1,500 horizontal ft, for a total 
combined distance of over 2,500 ft, or approximately one-half mile. Travel from stair exit and transfer floors increases this 
distance. 

NIST NCSTAR 1-7, WTC Investigation 149 



Chapter 10 



coworker wasn't feeling well and didn't think he/she could make it down 
the stairs. 

Occupants also observed other occupants in the stairs who were resting inside the stairwell. Face-to-face 
interview respondents sometimes mentioned specific categories of individuals who were resting in the 
stair, including people having trouble breathing (asthma), overweight, elderly, helpers, and firefighters. 

Leaving the Stairs 

In WTC 1 and WTC 2, occupants left the stairs for a variety of reasons. While occupants left the stairs 
onto floors throughout the buildings, the most commonly reported floor was one of the skylobbies (floor 
44 or 78). In addition, there were a number of occupants who left the stairs at the lower floors of WTC 1, 
which may be attributed to either debris from the collapse of WTC 2 or the fire department rest station for 
occupants somewhere between floors 12 and 20. 

In WTC 1, the most frequent reasons cited for leaving the stairs were an instruction to do so from 
firefighters, Port Authority, or building security (33 percent) and the stair condition (41 percent), 
including crowding or smoke/dust/jet fuel in the stairs. Before the collapse of WTC 2, occupants were 
directed out of certain stairs to either switch to another stair immediately or wait on the floor for a certain 
period of time. 

One occupant who originated on a low floor in WTC 1 stated that, "We 
were directed by a man (I am unsure if he was a security guard). He was 
telling people to go to the 18th floor in order to scatter the traffic from 
the stairwell. When we arrived at the 18th floor, we got out of the 
stairwell and into a vacant floor space." This occupant later stated that, 
"He said he was instructed to pile up the people into the vacant floors to 
control the flow of traffic." Interview 1000769 (NIST 2004) The 
occupant demanded to know why this was happening and decided to find 
another stair to take out of the building. A group then followed this 
occupant to another stairwell. 

After the collapse of WTC 2, the occupants near the lower part of WTC 1 were faced with an onslaught of 
debris from the collapse. For this reason, occupants were again instructed to leave the stairs and if not 
instructed, took it upon themselves to switch stairs at times. 

Two occupants helping an overweight colleague down Stair A in WTC 1 
after the collapse of WTC 2 "were told by firemen and rescue workers 
that [they] had to go up to the fourth floor [because] the exit was 
blocked." Interview 1000093 (NIST 2004) 

Another frequent reason for leaving the stairs in WTC 1 , which may have been unknowingly caused by 
either instruction or stair condition, was occupants following other occupants (9 percent). These people 
felt comfortable following others out of the stair, without always knowing the reason for leaving or who 
had initiated the stair move. Other reasons for leaving the stairs in WTC 1 include a jammed exit at the 
transfer floor (76) of Stair A, occupant's being uncomfortable, the presence of a mobility impairment, 
retrieving something, helping another person, and seeking information. 



1 50 NIST NCSTAR 1-7, WTC Investigation 



Discussion and Analysis 



In WTC 2, the most frequent reasons for leaving the stairs were instructions to leave the stairs by Port 
Authority or building security (10 percent) personnel, the public address system announcement made 
before WTC 2 impact (32 percent), and elevator usage (13 percent). Some occupants in WTC 2 were 
being instructed out of the stairwells at the skylobbies minutes before the announcement was made. 
When the announcement was made to occupants throughout the building, some occupants inside the stair 
walked out onto a floor to hear the announcement more clearly and/or to react to the information given by 
the announcement. Also, some occupants decided to take the elevators to a different point in the building 
(to either evacuate or return to their offices) both before and after the announcement was made. Other 
reasons for leaving the stairwell in WTC 2 included following the crowd, occupants being uncomfortable, 
and seeking information. 



10.3.7 



Aids and Constraints to Evacuation 



Evacuation incentives are interpersonal interactions or physical features of the built environment that 
helped people to evacuate. Analysis of the telephone interview data reveals that evacuees that among 
occupants who received help from other people, 9 percent reported that they were assisted by floor 
wardens, 44 percent reported that they were helped by police or firefighters, and 65 percent reported that 
they were assisted by "others." 

For occupants who were helped by building features, 33 percent of survivors in WTC 1 and 17 percent of 
those in WTC 2 reported that they were helped by photoluminescent markings. The discrepancy between 
towers may be due to the fact that lights were lost in WTC 1 after WTC 2 collapsed, thereby 
demonstrating the usefulness of the photo luminscent qualities. Additionally, occupants who used 
elevators in WTC 2 would not have observed the photoluminescent paint. 

Conversely, certain conditions presented constraints to evacuation. Table 10-1 shows the most frequently 
reported constraints to evacuation from the telephone interview data. Three areas were reported by more 
than half of the evacuees in WTC 1: crowded stairwells, emergency responders in the stairwells, and 
injured or disabled persons in the stairwells. The findings from the causal model for normalized stairwell 
evacuation time in WTC 1, however, provide a scientific basis for refuting the occupant's perception that 
firefighters entering the building adversely affected the overall flow down the stairwells. While more 
than half (63 percent) reported that the firefighters / police in the stairwells were a constraint, an occupant 
who encountered firefighter or police did not have a significantly slower stairwell travel time than an 
occupant who did not encounter firefighter or police in the stairwell, all other factors being held constant. 

Table 10-1. Constraints to evacuation. 





WTC 1 


WTC 2 


Stairwells were too crowded 


73 % (321) 


69 % (206) 


Firefighters/police in stairwell 


63 % (275) 


27 % (80) 


Injured/disabled in stairwell 


52 % (226) 


33 % (99) 


Lack of direction/information 


24% (104) 


29% (106) 


Locked doors 


16% (72) 


7 % (25) 


Poor lighting 


1 1 % (48) 


4% (15) 


Bad/missing signage 


5 % (23) 


5% (18) 



Source: NIST WTC telephone survey data. 



NISTNCSTAR 1-7, WTC Investigation 



151 



Chapter 10 



A possible explanation for this seeming inconsistency is that while an occupant was required to stop and 
stand to the side in the stairwell as firefighters and police moved past, they were able to temporarily 
increase their speed to catch back up to where they would have been before they had met the firefighters 
or police. This explanation is contingent upon the occupants descending the stairwell at less the 
maximum speed, which was found to be the case earlier in this chapter (the travel speed was about one- 
half what would be expected in a non-emergency evacuation). 

Elevator Usage in WTC 2 

At least 1 8 percent of the survivors from WTC 2 reported using the elevators for at least part of their 
evacuation, including those who used elevators from the basement levels. Those who used elevators in 
WTC 2 did so for various reasons. While most occupants used elevators for evacuation, some had 
decided to find a quick way to return to their office once the 9:00 a.m. announcement was given. Elevator 
usage was not necessarily dominated by people with mobility impairments, but used by all people with 
the intent to evacuate the building quickly. However, one occupant using the elevators out of need started 
out in the 90s of WTC 2. 

"I wanted to get out of the building as quickly as possible. I was taking a 
new medication and knew I should not walk down the stairs." Interview 
1000553 (NIST 2004) This occupant took an elevator from the 95th 
floor to the 78th and then switched to another elevator at the 78th floor to 
travel to the lobby level. 

Despite the availability of elevators for occupants with mobility impairments in WTC 2, however, 
approximately the same percentage of mobility-impaired occupants chose to use elevators in WTC 2 
(19 percent), when compared to the surviving population overall (16 percent). 

Face-to-face interview respondents also refer to elevator usage after the plane hit. An occupant, injured 
on a floor in the 70s, was evacuated via elevator by a firefighter and a security guard, along with two 
other injured occupants, after WTC 2 has hit by the airplane. 

"As we were walking down, we saw a fireman coming up and told us to 
get to 40 and that someone would take us in an elevator down to the 
lobby." Interview 1000562 (NIST 2004) 

The use of elevators in WTC 2 saved many lives due to the fact that occupants from floors 78 - 107 in 
WTC 2 used both stairwells and elevators in order to move below the impending impact region prior to 
the WTC 2 attack. In order to estimate the total number of occupants able to descend below the impact 
zone prior to the WTC 2 attack, the following assumptions were made: (a), no occupant began evacuating 
WTC 2 prior to 9:02:59 a.m.; (b), no elevators were usable; and (c), the evacuation rate of WTC 2 
mirrored the observed evacuation rate of WTC 1 in Figure 10-2 (starting at 8:46:30 a.m.). Under these 
three assumptions, over 3,000 people would have remained in the building as it collapsed, with over 
2,000 occupants remaining trapped above the 78th floor. Therefore, self evacuation (starting to evacuate 
prior to 9:02:59 a.m.) and the use of elevators during that time period is estimated to have saved roughly 
3,000 lives in WTC 2 on September 1 1, 2001. 



1 52 NIST NCSTAR 1-7, WTC Investigation 



Discussion and Analysis 



Footwear 

Evacuation speed on the stairs can be significantly affected by the choice of footwear worn by the 
occupants, especially women. High-heeled shoes, especially heels higher than 3 in., dramatically 
decrease movement capability on the stairs (Templer 1992). People, mostly women, removed their shoes 
in the stairwell in both towers. This presented a potential hazard for other occupants in the stair, who had 
to maneuver around the pile of shoes, as well as for the occupants without shoes walking through the 
damaged portions of the building. 

In WTC 1, an occupant noticed a pile of shoes in a stair near the 28th 
floor. "There was a pile of shoes that accumulated from people kicking 
them off. Some of the people around me were tripping on them and 
warning others to watch out for them." Interview 1000042 (NIST 2004) 

Firefighters (Interviews 1000081 and 1000540 (NIST 2004)) and Port Authority personnel (Interview 
1000071 (NIST 2004)) occasionally instructed occupants to put their shoes back on. 

Transfer Hallways 

As described in Chapter 2.2, the stairwells did not descend in a straight vertical alignment in WTC 1 and 
WTC 2. The horizontal connections, more numerous and lengthy in Stairwells A and C than Stairwell B, 
could extend more than 100 ft and require several 90 degree turns. In addition to slightly increasing 
evacuation time (compared to a design with no horizontal transfers), the transfer hallways introduced 
uncertainty in the minds of the evacuees regarding the correctness of their evacuation path. 

An occupant from a floor in the 60s in WTC1 described the transfer 
hallways: "As I descended the stairs down to the Mezzanine Level, once 
or twice I had to exit the staircase through a door and go down a corridor 
in order to reconnect to the same stairwell. I found this to be extremely 
disconcerting. Everyone who did this, stopped before they exited the 
staircase to make sure they were doing the right thing. This slowed us 
down and there was concern that the door would lock behind us." 
Interview 1000053 (NIST 2004) 

10.3.8 Emergency Responders and Building Authorities 

Emergency Responders 

In addition to organizing the response to the attacks on the two towers and assisting occupants during 
their evacuation, building staff and emergency responders had to use the stairwells to attempt access to 
impacted floors in the buildings. This resulted in small groups of firefighters in bulky bunker gear moving 
against the flow of occupants down the stairs. This phenomena is often referred to as counterflow. Police 
and fire department involvement with building occupants was identified by survivors as both an 
evacuation aid and constraint. The police and firefighters were identified as an evacuation aid by 
44 percent of the occupants in WTC 1 and 30 percent of the occupants in WTC 2 and as a constraint to 
evacuation by 62 percent of the occupants in WTC 1 and 27 percent of the occupants in WTC 2. The 
lower numbers should be expected since before WTC 2 was hit, emergency responders naturally 



NIST NCSTAR 1-7, WTC Investigation 1 53 



Chapter 10 



concentrated their efforts in WTC 1 . Many of those who evacuated early from WTC 2 never encountered 
police or firefighters in the building. 

Inside WTC 1 stairwells, firefighters were sighted by interviewees in all three stairwells, with a 
concentration on Stair B (from the face-to-face interviews). Also, from the face-to-face interviews, 
firefighters were sighted as high as the 60s in WTC 1. Interviews 1000576 and 1000645 (NIST 2004) For 
those meeting firefighters in the stairs, some mentioned slowing down, crowding, and even stopping 
several times. An occupant in stairwell A of WTC 1 was both slowed down and reassured by the 
firefighters. 

"We encountered firemen ascending, starting at about the 35th floor. I 
came to a dead stop numerous times in Stairwell A for about 5 minutes 
each time to aid the firemen to get up to the problem. The firemen were 
easy-going, and attempted to put people at-ease. They were also 
extremely winded and sweating profusely from their climb." Interview 
1000103 (NIST 2004) 

Figure 1 0-7 shows an FDNY firefighter ascending a 44 in. ( 1 . 1 m) stairwell in WTC 1 on 
September 11, 2001. Figure 10-8 shows how an occupant and a firefighter in bunker (turnout) gear may 
not be able to pass one other in a 1.1 m (44 in.) stairwell without either the occupant or the firefighter 
moving somewhat to the side. Figure 10-8 was not taken in a stairwell from the WTC complex, but was 
intended to be a generic demonstration of the counterflow phenomena. 




Figure 10-7. Firefighter and occupants using a 44 in. stairwell in WTC 1 on 

September 11, 2001. 



154 



NIST NCSTAR 1-7, WTC Investigation 



Discussion and Analysis 



The firefighters provided a sense of 
security to some occupants in the towers. 
In some cases, occupants went so far as to 
show their appreciation. An occupant 
from WTC 1 Stairwell A recalled 
encountering firefighters near the 20th 
floor. 

"Firemen were going up and they were 
looking exhausted. People were 
cheering the firemen and some of the 
firemen said, 'Don't stop, you're 
almost there.' As we moved aside, we 
handed the firemen water." Interview 
1000722 (NIST2004) 

Throughout the event, emergency 
responders supported the occupants by 
performing a variety of actions, including 
directing occupants to change stairs 
(especially in WTC 1 when exits were 
blocked by debris from the collapse of 
WTC 2), providing directions on how to 
exit through the Concourse, giving out 
snacks, water, and oxygen (from air 
bottles) to occupants in need. 















■* *MI 




^d 


w 


A fH 




II 


- 











Source: NIST. 



Figure 10-8. Occupant and firefighter in bunker 
gear passing in a generic 44 in. stairwell. 



Firefighters reportedly established a rest 

station somewhere between floors 12 and 

20 (Interview 1000543 [NIST 2004]). Firefighters here instructed occupants to drop certain occupants off 

at this floor for assistance, as well as suggesting that evacuees stop on this floor if they required rest or 

assistance. 



In addition to walking up toward the fire floors, many firefighters moved injured occupants to safety. An 
occupant inside Stairwell C of WTC 1 heard instruction from floor above to keep to the right. 

The respondent then "noticed that the firemen in uniforms were bringing at least two people" down the 
stairs. "One gentleman had his arms severely burned and one lady who had her head halfway burned and 
was screaming with pain. After the fireman came down, others came up - several of them. They had 
gear on their back and each one had an extra hose on their shoulder. I was on the right and was letting 
them up, and a particular firemen patted me on the left shoulder and said, 'Don't worry, you will be ok.'" 
Interview 1000697 (NIST 2004) 



NISTNCSTAR 1-7, WTC Investigation 



155 



Chapter 10 



Impact of Authorities on Occupant Evacuation 

In both towers, occupants followed directions from those in a position of authority. In an office 
emergency, higher rank can mean multiple things. On the floors, occupants followed instructions to leave 
from supervisors or fire wardens, even though supervisors and wardens may have had the same 
information as the employee about what was going on. Also, occupants followed directions throughout 
the building from the building and fire officials who were familiar with the building layout, such as the 
Port Authority employees, police, firefighters, and building security. Analysis of face-to-face interview 
data revealed that these instructions included when and where to evacuate the stairs, when and where to 
change stairs, when to move to the right to let firefighters go up and injured down, or where to travel upon 
leaving the stairwell. 

Workplace Authority 

In an emergency, there is usually a combination of reasons why a person begins their evacuation 
process. 5 However, there is usually one significant reason that finally makes them decide to leave or that 
weighs more heavily on their decision than the others. When occupants in both towers were asked to 
name the one thing that made them decide to evacuate, 14 percent of the occupants in WTC 1 and 
21 percent of the occupants in WTC 2 said that the reason they evacuated was being told to evacuate. 
(Telephone Interviews, NIST 2004) 

At the first awareness that something was wrong, the occupants could only rely on each other (or 
themselves, if alone) to understand what had happened and to decide what to do next. Fire alarms 
sounded in certain areas of WTC 1 (and possibly WTC 2); however, occupants did not generally report 
the fire alarms as their reason to evacuate. The others on their floor were their subordinates, coworkers, 
or superiors (supervisors or fire wardens). From a total of 208 face-to-face interviews, 86 people 
mentioned being instructed to evacuate, as well as their reaction to that instruction. Most respondents 
began their evacuation when told to leave, whether by a superior or co-worker. Although the percentage 
was higher when direction came from a superior, it is difficult to draw any conclusion from the face-to- 
face data since it is not statistically -based and cannot be generalized to the entire population of the 
buildings. Only one face-to-face interview respondent reported telling their superior to evacuate, and in 
that isolated case, the superior ignored the instruction. 

An occupant from the 40th floor in WTC 1 left the floor due to persuasion by the fire warden: 

"As I was typing the email message, I heard a loud voice say 'Leave.' I 
kept typing the message when a fire warden grabbed my arm and pulled 
me out of the seat." The fire warden also pointed this occupant in the 
direction of the elevators and stairs in addition to instructing the 
occupant to leave the floor. Interview 1000802 (NIST 2004) 

After hearing an explosion from WTC 1 , an occupant with workplace authority on a floor in the 60s of 
WTC 2 instructed others on the floor to leave: 



Nelson and MacLennan. "Emergency Movement." In The SFPE Handbook of Fire Protection Engineering, 2 n Edition. 
NFPA, Quincy, MA. 1995. 

156 NIST NCSTAR 1 - 7, WTC Investigation 



Discussion and Analysis 



"I saw fire and burning debris and smelled jet fuel. I ran out of my office 
and yelled to staff and co-workers to order them to evacuate. We ran to 
the stairwell." Interview 1000556 (NIST 2004) 

And, even after WTC 2 was hit, occupants on a floor in the 50s of WTC 2 remained on their floor 
discussing what had just happened and what they should do next: 

"Our facilities manager told me it was time to go. We did a quick check 
of the floor to make sure no one was left behind. In the hallways we saw 
that everyone was heading in one direction and we followed them to the 
stair." Interview 1000557 (NIST 2004) 

It would appear from the face-to-face interviews that occupants were likely to follow the instruction given 
by their superior in an emergency. However, even if the final decision to evacuate was prompted by the 
instruction to evacuate, other factors may have been involved in making that decision, including seeing 
the plane strike into WTC 1, past experience, or other observations of the event. 

Building Authority 

Similar to the FDNY role in the WTC towers on September 11, 2001, the building authority played an 
important role in providing guidance to occupants about where to go once they left their floors, which the 
occupants frequently followed. The Port Authority personnel were observed giving instructions and 
directions to occupants at the skylobbies, Mezzanine, and Concourse areas of the WTC towers. Their 
instructions for some occupants consisted of when and where to leave the stairs, whether to use the 
elevators, and in WTC 2, when to return to their offices. Port Authority direction consisted of how and 
where to go through the building and Concourse area in order to leave the WTC complex. In most cases, 
occupants followed the instructions and welcomed the directions given by the building authority in the 
towers. 

Many noted the tremendous help that building personnel provided at the base of the building. An 
occupant in WTC 2 noted that the building authorities were present throughout the Concourse: 

"Trade center people were directing us into the Concourse because you 
couldn't go out Liberty street - they had all those exits closed because 
there was debris flying all over the street. Security guards were like a 
human chain telling us which direction to go. We followed the security 
guards' direction . . . towards Borders." Interview 1000842 (NIST 2004) 

An occupant in WTC 1 noted the Port Authority giving directions to occupants on the Mezzanine: 

"As soon as I arrived onto the Mezzanine floor, I saw a chief operating 
officer giving directions to get onto the stairs [escalator] and go down. 
He also said not to look out onto the plaza because it was unsafe and 
dangerous." Interview 1000639 (NIST 2004) 

In addition to directing occupants throughout the buildings, there were two cases from the face-to-face 
interviews where occupants were rescued from their floors by the help of the building authority in 



NIST NCSTAR 1-7, WTC Investigation 1 57 



Chapter 10 



WTC 1. In both cases, the occupants were faced with worsening conditions on their floor and were 
guided to the stairwells by the official. 

From a floor in the 80s in WTC 1 : After preparing their office with wet 
towels around the doors to prevent the heavy smoke from seeping in, the 
respondent "saw a beam of light through the smoke, which turned out to 
be a fireman and a building worker. They told us to drop everything and 
follow them into the hall that led to the staircase." Interview 1000055 
(NIST 2004) 

After two unsuccessful attempts to open the jammed door from a floor in the 80s into stairwell A, the 
heavy smoke on the floor drove the group of coworkers back to their office to wait, until: 

"We heard one of our coworkers in the hall [say that] the Port Authority 
fellow had opened up the doors to Stairwell A for us." Interview 
1000137 (NIST 2004) 

10.3.9 Occupant Experience 

Experience from the 1993 bombing and other past evacuation experience affected some occupants' 
actions in 2001. For example, many of those involved in both 1993 and 2001 recalled long evacuation 
times in 1993, including occupant congestion, smoke in the stairs, and not being allowed to return to the 
building for weeks after the event. Because of this experience, some occupants started their evacuation as 
soon as WTC 1 was hit and noted that their experience in 1993 was the principal reason. On the other 
hand, some occupants performed specific activities that they wished they had performed in 1993, such as 
calling home so their parents/family wouldn't worry about them and taking certain belongings in case 
they couldn't return to the building right away. 

Even though occupants of the towers may not have been present for the 1993 bombing, the bombing 
event may have played a role in their 2001 evacuation. Many of those present in 1993 shared their 
experiences with other employees both before and during the 2001 attack: 

"I was not in the building in 1993, but a lot of people who had been there 
during the 1993 bombing were very helpful because they were exiting 
faster. They knew that every second counted based on that previous 
experience and I took my cue from them." Interview 1000518 (NIST 
2004) 

"I wasn't at the building in 1993, but I knew that it took over an hour to 
get down the stairs, which influenced my decision to use the elevators." 
Interview 1 00073 1 (NIST 2004) 

10.3.10 Mobility-Impaired Occupants 

When evacuating a high-rise building, many different physical and medical conditions can affect travel 
ability on stairs. As the total distance traveled to reach an exit increases, the number of people unable to 
successfully complete the evacuation without resting or requiring assistance increases. In the WTC, stair 
travel challenged occupants with certain conditions, such as wheelchair use, pregnancy, asthma, visual 



158 NIST NCSTAR 1-7, WTC Investigation 



Discussion and Analysis 



impairment, physical impairment, obesity, arthritis, and old age. The current NYC Building Code defines 
a physical disability as any one of the following: 

• Impairment requiring the use of a wheelchair; or 

• Impairment causing difficulty or insecurity in walking or climbing stairs or requiring the use of 
braces, crutches, or other artificial supports; or 

• Impairment caused by amputation, arthritis, spastic condition or pulmonary, cardiac, or other ills 
rendering the individual semi-ambulatory; or 

• Total or partial impairment of hearing or sight causing insecurity or likelihood of exposure to 
danger in public places; or 

• Impairment due to conditions of aging. 

Evacuation of WTC 1 and WTC 2 during the 1 993 bombing identified an inefficiency at the World Trade 
Center in evacuating the mobility-impaired. The Report of the World Trade Center Review Committee in 
1995 indicated: "Evacuation of persons with disability from the World Trade Center was slow and 
arduous, the [sic] Committee proposes that methods of evacuation should be studied to provide equally 
safe egress for these building occupants." (New York City 1995) 

Despite introduction of evacuation chairs and a buddy system for pre-identified mobility-impaired 
occupants, 51 percent of the occupants in WTC 1 and 33 percent of the occupants in WTC 2 in 2001, 
noted that injured and disabled persons in the stairwell were a constraint to evacuation. However, 
occupants were quick to aid these individuals by guiding them throughout their evacuation or simply 
moving to the side of the stairwell to let those who were injured and others in need pass by when they 
could. 

In WTC 1, "someone was being carried down in some kind of 
handicapped apparatus and was strapped in. The occupant was being 
carried by two fellow occupants around the 30s. We stopped to allow 
them access." Interview 1000834 (NIST 2004) 

In some cases, occupants noted passing slower mobility-impaired individuals in the stairs and even 
slowing or stopping behind them. 

In WTC 2, "we saw an [occupant] who was hyperventilating. [The 
occupant] was walking down the stairs with assistance. We slowed 
down and came to a stop [because] we couldn't get around the two 
[occupants]." Interview 1000556 (NIST 2004) 

In WTC 1, two occupants were helping an overweight occupant evacuate 
the building. During one of the helper's descent down the stairs, the 
helper noted that "we took up the entire width of the stairway and no one 
could get around us until we came to a landing." Interview 1000093 
(NIST 2004) 



NIST NCSTAR 1-7, WTC Investigation 1 59 



Chapter 10 



Finally, some occupants reported mobility-impaired occupants waiting on the stairs and/or landings for 
others to help them or to be rescued by the fire department. Many respondents recalled two specific 
occupants from WTC 1 who weren't able to evacuate the building in time. These included an occupant in 
a wheelchair waiting with a friend on the stairs and an occupant with arthritis waiting at a fire department 
rest stop somewhere between floors 12 and 20. 

Mobility-Impaired Occupants and Mortality below the 92nd floor in WTC 1 

Several occupants perished assisting mobility-impaired colleagues (Fahy and Proulx 2003) and many 
more occupants and emergency responders risked their lives assisting mobility-impaired colleagues who 
successfully evacuated. However, for occupants where a likely mechanism contributing to unsuccessful 
evacuation could be identified, being trapped by debris on the starting floor, delayed evacuation initiation 
(of statistical outlier magnitude), or performing emergency response building responsibilities accounts for 
the majority of the below the impact region deaths. 

10.4 EVACUATION SIMULATIONS 

The purpose of modeling evacuation from the World Trade Center towers was to obtain evacuation times 
for a variety of scenarios in order to provide additional context with which to understand the 
September 11, 2001 evacuation of WTC 1 and WTC 2. Table 10-2 shows each of the six general egress 
simulations performed, along with details regarding the evacuation type (full building evacuation or 
phased evacuation, also known as defend-in-place), the number of evacuees included in the simulation, 
the input response delay, which models were used, and any other relevant information about the 
simulation. There were five full building evacuations and one phased evacuation simulated. The full 
building evacuation simulations explored the effect on the total evacuation time (or number of successful 
evacuees if the time was fixed) with respect to the number of simulated evacuees, the presence or absence 
of building damage (observed on September 11, 2001), and the type of model used for the simulations. It 
should be noted that none of the models used in this analysis have been validated for emergency 
evacuation in buildings as large as 1 10 stories. 

10.4.1 Egress Simulation Results 

Phased Evacuation 

The following section is a summary of the egress simulation results. For a more complete discussion of 
egress modeling inputs, assumptions, limitations, and results, refer to Appendix D: Egress Modeling. The 
purpose of simulation 1 , phased evacuation, was to understand not only the time necessary to perform a 
phased evacuation, but to compare the results using three different egress models: Simulex, EXIT89, and 
buildingEXODUS. 

Table 10-3 shows the total phased evacuation times for each model. The three models simulate a total 
phased evacuation time between 3!/2 and 4 minutes for all 600 occupants, assuming that evacuees start 
evacuating immediately. If the evacuees were randomly assigned a delay time between and 10 minutes, 
the simulated total evacuation time was between 1 1 and 1 1 14 min for all 600 occupants. Phased 
evacuation is an efficient strategy to quickly remove occupants most at risk quickly from 'routine' 
emergencies, or those that fit within the design envelope of the life safety systems. 



160 NIST NCSTAR 1-7, WTC Investigation 



Discussion and Analysis 



Table 10-2. Egress simulation matrix. 



Scenario 


Full Building or 

Phased 

Evacuation 


Number of 
Evacuees 
Simulated 


Response 
Delay 


Model(s) 
Used 


Notes 


1 


Phased 


600 


Zero 

- 10 min 


Simulex 

EXODUS 

EXIT89 


• Occupants travel three floors 
below fire floor 


2 


Full Building 


19,800 


Zero 

0- 10 min 


EXODUS 
EXIT89 


• Fully-occupied, without visitors 

• No damage 


3 


Full Building 


25,000 


Zero 

- 10 min 


EXODUS 
EXIT89 


• Fully-occupied, with visitors 

• No damage 


4 


Full Building 


8,800 


Zero 

- 10 min 


EXODUS 
EXIT89 


• September 11, 2001 population 

• No damage 


5a 


Calibration 
Simulation 


7,200 


6-30 min 


EXODUS 


• Stop and Go 

• Ground - Floor 90 

• Damage above Floor 90 


5b 


FullWTC 1, 
with damage 


16,000 


6-30 min 


EXODUS 


• Stop and Go 

• Ground - Floor 90 

• Damage above Floor 90 


6a 


Calibration 
Simulation 


7,400 


2-17 min 
(Above 
Floor 77) 
6-30 min 
(Ground to 
Floor 76) 


EXODUS 


• Stop and Go 

• Ground - Floor 76 

• No Damage 


6b 


Elevator, WTC 2 


19,800 


Zero 


ELVAC 


• 14 Minute Elevator Simulation 


6c 


Full WTC 2, 
With Damage 


17,260 


2-17 min 
(Above 
Floor 77) 
6-30 min 
(Ground to 
Floor 76) 


EXODUS 


• Stop and Go 

• Ground to Floor 107 for first 16 
min 

• Ground to Floor 77 for time range 
1 7 min - 72 min 



Ta 


Die 10-3. Results for phased evacuation simulations. 




Occupant Type, 


Evacuation Time (s) 


Evacuation Model 


Characteristics 


No Delay 


0-10 min Delay 


Simulex 


All office staff 
60 % men 
40% women 


240 


690 


EXODUS 


5 % males, age 17-29 
38% males, age 30 -50 

21 % males, age 51 -80 
3 % females, age 17-29 

22 % females, age 30 - 50 
11 % females, age 51 - 80 


243 


660 


EXIT89 


All medium body size 
Emergency speed 


210 


690 



NISTNCSTAR 1-7, WTC Investigation 



161 



Chapter 10 



Full Building Evacuation 

The purpose of simulations 2, 3, and 4 was to explore the effect of additional building occupants on the 
total building evacuation time. The three populations selected were (1) the NIST estimate of building 
population on September 11, 2001, (2) the NIST estimate of a fully-occupied tower without visitors 
(19,800 occupants), and (3) the NIST estimate of a fully-occupied tower including visitors 
(25,500 occupants). Note that in 2005, Port Authority estimated that the maximum population of WTC 1 
or WTC 2 would not likely have exceeded 20, 000. 66 Figure 10-9 shows the model results for the three 
building populations using the results from the buildingEXODUS model. The slope of the regression was 
identical whether comparing the September 11, 2001 population evacuation time to fully-occupied 
building evacuation time ([112 min - 55 min] / (19,800 - 8,800) = 5.2 min per thousand additional 
occupants) or compared to fully-occupied with visitors building evacuation time ([142 min - 55 min] / 
(25,600 - 8,800) = 5.2 min per thousand additional occupants). Thus, for a given building geometry, and 
a number of other important simulation assumptions, adding additional occupants to a building population 
resulted in a linear increase in total building evacuation time. 

Extrapolation of Evacuation Time for a Fully-Occupied Tower on September 11, 2001 

In Chapter 4.1 of this report, NIST estimated that the number of successful evacuees from WTC 1 on 
September 11, 2001 was 7,500, the total of which required roughly 100 minutes to exit the building. As 
shown in Figure 10-9 and Table 10-4, the buildingEXODUS model estimated that a fully-occupied 
building (population 25,500) required approximately 2.6 times as long (142 / 55 = 2.6) as a building with 
a September 11, 2001 occupant load; therefore, on September 11, 2001, a fully-occupied WTC tower with 
visitors, may have required roughly (100 x 2.6) = 260 min (over 4 hours) to fully evacuate. 

Using the same logical approach, if WTC 1 had been occupied by approximately 20,000 occupants on 
September 11, 2001 (fully-occupied without visitors), the evacuation would have taken (1 12 min / 55 min 
= 2.0 x 100 min = 200 min) over 3 hours to complete. 



Table 10-4. Total building evacuation time (simulated) for various occupant loads. 


Evacuation Model 


Evacuation 

Initiation Delay 

Input 


Evacuation Time 

(min): 8,800 

occupants 


Evacuation time 

(min): 19,800 

occupants 


Evacuation 
time (min): 

25,500 
occupants 


EXODUS 


10 Minute Delay 


55 


112 


142 


No Delay 


52 


110 


141 


EXIT89 


10 Minute Delay 


71-74 


92-113 


119-139 


No Delay 


58-78 


97-117 67 


114-140 



66 Bhol, Saroj. PANYNJ (September 21, 2005). Email from S. Bhol to S. Sunder in response to NIST question. 

The underlying theory for people movement in EXIT89 is based upon the work of Predtechenskii and Milinskii. They 
observed an inverse relationship between density and speed for three different types of movement: emergency, normal, and 
comfortable. The effect of crowding (density) on overall evacuation speed was greater when no delay time was assumed. A 
delay time when 19,800 occupants were present, according to EXIT89, spaces the occupants out and increases overall 
evacuation efficiency. This was not found to be the case for all occupant loads (25,500). 



162 



NIST NCSTAR 1-7, WTC Investigation 



Discussion and Analysis 



Full Building Evacuation of a WTC Tower 



160 

140 

"c 120 

'E 

I* 100 

80 



c 
o 

3 

o 

ill 40 



60 



20 















A 


25,585 - full capacity 
with visitors 




















19,800 - full capacity 
























8,800 - 9/11 capacity 




















III 





5000 10000 15000 20000 

Number of occupants in Tower 



25000 



30000 



Figure 10-9. Simulation of full building evacuation of a WTC tower with different 

occupant loads. 

Estimated Occupant Mortality from a Fully-Occupied WTC Tower on September 11, 2001 

Scenarios 5 and 6 were simulated in order to estimate the consequences of a fully-occupied, without 
visitors (total building population: 19,800) from WTC 1 or WTC 2 on September 11. Each tower was 
simulated with a September 1 1 population, and model assumptions and inputs (such as evacuation 
initiation delay time) were refined in order to roughly match the actual outcome on September 11. After 
the buildingEXODUS model results were determined to resemble gross characteristics of the 
September 1 1 evacuation, more occupants (fully-occupied building, without visitors) were added, and the 
results are presented below. 

WTC 1 was assumed to have no passable stairwell above floor 91 starting at time zero. Elevators were 
assumed to be rendered inoperable. The simulation was ended at 1 03 minutes and any simulated 
occupant remaining in WTC 1 was assumed to have perished. The simulation shows that 69 percent of all 
occupants (13,600 occupants out of 19,800) would have evacuated WTC 1 prior to collapse: percent 
(0 out of 3,800) from above the impact zone and 85 percent (13,600 out of 16,000) from below the impact 
zone. Table 10-5 summarizes the results of scenarios 5 and 6. 



NISTNCSTAR 1-7, WTC Investigation 



163 



Chapter 10 



Tabl 


e 10-5. Simulated evacuation results for fully- 

September 11, 2001. 


occupied WTC 1 and WTC 2 on 


Building 


Total 

Number of 

Occupants 

at t = 0.0 


Potential 
Number 

of 
Evacuees 


Occupants 
Remaining 
in Building 
at Collapse 


Occupants 
Trapped 

Above 

Floors of 

Impact 


Percentage of Occupants Who 

Successfully Evacuated Relative to 

Where They Started 


Total 


Below 
Impact 


Above 
Impact 


WTC 1 


19,800 


16,000 


6,200 


3,800 


69% 


85% 


0% 


WTC 2 


19,800 


17,260 


8,377 


3,900 


58% 


75% 


44% 


Total 


39,600 


33,260 


14,577 


7,700 


63% 







Occupants of WTC 2 were assumed to use elevators for a period of 1 6 minutes, after which all elevators 
were assumed inoperable. After 16 minutes, when WTC 2 was attacked, it was assumed that no 
occupants survived above floor 77. The simulation was ended after 72 minutes, when WTC 2 collapsed. 
The simulation showed that 1 1,423 of the 19,800 occupants, or 58 percent of the initial population, would 
have successfully evacuated. Of the 1 1,423 simulated occupants, 8,883 evacuated using the stairwells, 
while 2,540 simulated occupants evacuated using elevators. Of the 8,377 occupants who remained in the 
building at the time of collapse, 3,900 simulated occupants were trapped above floor 77. Another 4,477 
were evacuating from below the 78th floor when WTC 2 collapsed: 1,231 simulated occupants were 
initially above the 77th floor, while 3,246 did not evacuate despite starting on a floor below the 
78th floor. The initial population of occupants below the 78th floor was 12,783. Therefore, (3,246 / 
12,783) 25 percent of the occupants who started below the impact floors did not successfully evacuate. 
Similarly, 3,900 of the initial 7,017, or 56 percent of the occupants at or above the impact region at 
8:46:30 a.m. in WTC 2 did not successfully evacuate. The simulation showed that 549 occupants initially 
above floor 77 successfully evacuated using stairwells, while only 3 of the 549 simulated occupants 
originated above floor 100. 

When combining the evacuation modeling results for WTC 1 and WTC 2, assuming 19,800 occupants in 
each tower and assuming that the aircraft impacts and collapses occurred at the same times as they did on 
September 11, 2001, about 14,000 occupants may have perished, not including any emergency 
responders, aircraft passengers, or bystanders. 

10.4.2 Egress Modeling Conclusions 

The egress modeling revealed three principal insights into the evacuation of WTC 1 and WTC 2 on 
September 11, 2001. First, while not an appropriate strategy for an emergency the scale of 
September 11, 2001, phased evacuation, under certain circumstances, moves occupants most at risk to a 
place of relative safety much more quickly and with less total impact upon building tenants than full 
building evacuation. Second, additional occupant load in a WTC tower (compared to September 11, 
2001) resulted in a linear increase in total building evacuation time. Third, assuming that the ratio of 
observed to simulated evacuation time of 2.6 would extend to the fully-occupied with visitors 
(25,500 occupants), full building evacuation scenario, WTC 1 would have taken over four hours to fully 
evacuate under the conditions of September 11, 2001. 



164 



NIST NCSTAR 1-7, WTC Investigation 



Chapter 11 

Findings 



The evacuations of World Trade Center (WTC) 1 and WTC 2 on September 11, 2001, were documented 
and analyzed. In order to provide context for the 2001 evacuations, relevant historical egress events were 
explored. The evolution of the egress and communication systems, as well as the emergency procedures 
within WTC 1 and WTC 2 was documented. Over 1 ,000 interviews, using advanced interrogatory 
methods, were conducted. Hundreds of published accounts from a variety of sources were collected and 
analyzed. Emergency call records, emergency communication transcripts, Occupational Safety and 
Health Administration complaints, and Port Authority of New York and New Jersey design records and 
plans were analyzed. Analysis and compilation resulted in the following conclusions. 

11.1 BUILDING POPULATION AND DEMOGRAPHICS 

• There were 8,900 ± 750 people in WTC 1 at 8:46:30 a.m. on September 11, 2001. Of those, 
7,470 (or 84 percent) survived the attacks, while 1,462 - 1,533 occupants were killed in WTC 1. 
The National Institute of Standards and Technology (NIST) Investigation found that at least 107 
occupants were killed below the 92nd floor. 



• 



• 



There were approximately 8,540 ± 920 people in WTC 2 at 8:46:30 a.m. on September 11, 2001. 
Of those, 7,940 (or 93 percent) survived the attacks, while 630 - 701 occupants were killed in 
WTC 2. The NIST Investigation found that at least 1 1 occupants were killed below the 78th 
floor. 

Sixty-seven percent of WTC 1 occupants and 51 percent of WTC 2 occupants had started 
working at the World Trade Center during the four years prior to September 11, 2001. 

Two-thirds of WTC 1 and WTC 2 occupants participated in at least one fire drill in the twelve 
months prior to September 11, 2001. Seventeen percent of the occupants did not participate in a 
fire drill during that time period, and 17 percent did not remember whether they had participated 
in a fire drill during that time period. 

• Nearly all occupants (93 percent) who participated in a fire drill were instructed about the 
location of the nearest stairwell. 

11.2 EVACUATION 

• Approximately 87 percent of WTC occupants, and over 99 percent of those below the floors of 
impact, were able to successfully evacuate. 

- At 9:02:59 a.m., when WTC 2 was hit and 17 minutes after WTC 1 was hit, 21 percent of 
survivors had exited WTC 1 and 41 percent of survivors had exited WTC 2. 



NIST NCSTAR 1-7, WTC Investigation 1 65 



Chapter 11 



- By 9:37 a.m., 22 minutes before collapse, 95 percent of survivors had exited WTC 2. 

- At 9:58:59 a.m., when WTC 2 collapsed, 88 percent of survivors had exited WTC 1. 

- By 10:12 a.m., 16 minutes before collapse, 95 percent of survivors had exited WTC 1. 

• Occupants of WTC 1 overwhelmingly initiated their own evacuation. The NIST Investigation 
found no evidence that public address system announcements were heard by occupants of the 
building, although the fire command station was attempting to make announcements. 

• Self-evacuation and use of elevators for 16 minutes in WTC 2 saved roughly 3,000 lives. 

• During the last 20 minutes before each building collapsed, the evacuation rate in both buildings 
had slowed to about one-fifth the immediately prior evacuation rate. This suggests that for those 
seeking and able to reach and use the undamaged exits and stairways, the egress capacity (the 
number and width of exits and stairways) was adequate to accommodate survivors. 

• In WTC 1, the average surviving occupant spent approximately 48 seconds per floor in the 
stairwell. That does not include any time prior to entering the stairwell, which was often 
substantial. 

• Some occupants of WTC 1 delayed or interrupted their evacuation resulting in over 100 deaths 
below the impact region. 

• The NIST Investigation found no evidence to indicate that anyone who was above the 91st floor 
in WTC 1 after the airplane impact survived. This was due to the fact that the stairwells and 
elevators were destroyed and helicopter rescue, despite several attempts by both occupants and 
aircraft, was not possible. 

• In WTC 2, approximately 75 percent of the occupants above the 78th floor at 8:46:30 a.m. had 
successfully descended below the 78th floor prior to the airplane impact at 9:02:59 a.m. This 
occurred in spite of conflicting announcements, first urging people to return to their offices 
around 9:00 a.m., and then informing them that they may initiate an evacuation if conditions 
warranted around 9:02 a.m. 



• 



• 



The NIST Investigation found evidence that 18 people successfully used Stairwell A in WTC 2 to 
leave the building after being on or above the 78th floor when Flight 175 hit the building. One of 
the 18 later died from injuries sustained on September 11, 2001. Additionally, at least two people 
went to or above the 78th floor after having been below the 78th floor at the time of impact in 
order to assist trapped people and did not survive the collapse of WTC 2. 

Minutes prior to the collapse of WTC 2, an NYPD Emergency Services Unit (ESU) officer 
radioed from a floor in the 20s to the outside that he was having trouble ascending the stairwell 
due to the large number of occupants descending (Interview 24 NYPD [NIST 2004]). While the 
origin of the occupants remains unknown, only 1 1 occupants who started evacuating below the 
impact region were known not to have survived. 



1 66 NIST NCSTAR 1-7, WTC Investigation 



Findings 



• 



Computer egress modeling indicated that a full capacity evacuation of a single WTC tower with 
approximately 20,000 occupants required a minimum evacuation time of 1 h and 50 min. Given 
that the actual evacuation time on September 11, 2001, was about 100 min without elevator use, a 
full capacity evacuation of the WTC towers by 20,000 people would likely have required 
somewhat greater than 3 h (2 times 100 min). To achieve a significantly faster total evacuation at 
full capacity would have required increases in egress capacity (number and width of exits and 
stairways). 

Egress modeling indicates that, had WTC 1 and WTC 2 been fully-occupied on September 11, 
2001, with approximately 20,000 occupants each, about 14,000 occupants may have perished in 
the building collapses. 

11.3 DELAYS IN EVACUATION 



• 



• 



Occupants in WTC 1 delayed starting their evacuation because environmental cues (information 
from the physical environment that something was terribly wrong) and floor (increased distance 
to safety) caused people to seek additional information. Next, the act of seeking additional 
information, that is "milling" about to make sense out of the situation, led people to take actions 
to prepare to evacuate. Finally, taking those actions to prepare to evacuate delayed the initiation 
of actually evacuating. Although there were other factors that had lesser influence on what 
people did, the paths of causal influence that defined the main process that led people to delay in 
the evacuation of WTC 2 on September 11, 2001 were identical to those for WTC 1, with one 
decided difference. Perceived risk was predicted by environmental cues and initial floor and also 
contributed to seeking additional information and taking pre-evacuation actions in WTC 2, while 
the effect of perceived risk was substantially lower in WTC 1. 

• Starting floor (increased distance to safety) substantially increased the odds that people would 
encounter environmental cues. Floor also increased delay in starting evacuation (this relationship 
is elaborated upon in much greater detail in Chapter 10 of this report), which, in turn, also 
increased the chances that people would encounter environmental cues. Observation of 
environmental cues (information from the physical environment that something was terribly 
wrong) in the stairwell had a large and direct effect on increasing the amount of time that people 
spent, on average, in their evacuation stairwell. In addition to this multi-step process with 
environmental cues as the key predicting variable, interrupting the process of evacuation for any 
reason increased the amount of time, on average, that people used to descend their evacuation 
stairwell. 

• Contrary to the perceptions of the occupants, counterflow in WTC 1 was determined by causal 
modeling analysis not to be a significant predictor of increased total evacuation time while in the 
stairwells when compared to other factors, including evacuation interruption and environmental 
cues. 



• 



Occupants in WTC 1 delayed starting their evacuation because environmental cues (information 
from the physical environment that something was terribly wrong) and floor (increased distance 
to safety) caused people out to find additional information, most likely information about what 
was going on and what they should do about. Next, the act of seeking additional information, that 



NIST NCSTAR 1-7, WTC Investigation 1 67 



Chapter 11 



is "milling" about to make sense out of the situation, led people to take actions to prepare to 
evacuate. Finally, taking those actions to prepare to evacuate delayed the initiation of actually 
evacuating. Although there were other factors that had lesser influence on what people did, the 
main paths of causal influence that defined the main process that led people to delay in the 
evacuation of WTC 2 on September 1 1th were identical to WTC 1 with one decided difference. 
This was that perceived risk was predicted by environmental cues and initial floor and also 
contributed to seeking additional information and taking pre-evacuation actions in WTC 2 while 
the effect of perceived risk was substantially lower in WTC 1. This was likely the case because 
WTC 1 was hit without warning, and only the people in WTC 2 had time to wonder (perceive) if 
their tower was going to be a target. 

11.4 EMERGENCY MANAGEMENT AND PLANNING 

• Building occupants, 9-1-1 operators, fire department dispatch, WTC building officials, and Port 
Authority personnel lacked adequate situational awareness despite nearly constant cross- 
communications. Many opportunities to communicate important information in a timely manner 
were missed, such as telling building occupants the general location of the impact region or 
whether to evacuate or not. 

• Faced with an uncertain situation, occupants of both buildings received conflicting feedback / 
advice from a variety of sources (including 9-1-1 operators, FDNY, family and friends, and the 
Port Authority) regarding whether to evacuate, whether to break windows, and what the nature of 
their situation was. 

• World Trade Center occupants were inadequately prepared to encounter horizontal transfers 
during the evacuation process and were occasionally delayed by confusion as to whether the 
hallway led to a stairwell and confusion about whether the transfer hallway doors would open or 
be locked. 

In addition to an announcement in WTC 2 just prior to the airplane impact, announcements were 
made from the fire command station in the lobby of WTC 2 after the aircraft impact, although the 
NIST Investigation found no evidence that any surviving occupants heard these announcements. 

The decision to establish the primary evacuation route underground through the Concourse (mall) 
and out up to street level by WTC 5 (commonly recalled as being by the Borders Bookstore) 
prevented a significant number of injuries and/or deaths. 

The first "first responders" were colleagues and regular building occupants. Acts of everyday 
heroism saved many people whom traditional emergency responders would have been unable to 
reach in time. 

11.5 OCCUPANTS WITH MOBILITY IMPAIRMENTS 



• 



• 



• 



Approximately 1,000 surviving occupants (a projection of the 6 percent reported in the telephone 
interviews) of WTC 1 and WTC 2 had a limitation that impacted their ability to evacuate, 
including recent surgery or injury, obesity, heart condition, asthma, elderly or otherwise requiring 



1 68 NIST NCSTAR 1-7, WTC Investigation 



• 



Findings 



assistance to walk, pregnancy, and others. The most frequently reported disabilities were recent 
injuries and chronic illnesses; the number of occupants requiring use of a wheelchair was very 
small, relative to the frequency of other mobility impairments. 

While many mobility-impaired individuals were able to successfully evacuate, often with 
assistance from co-workers or emergency responders, others were temporarily removed from the 
stairwells in order to allow more able occupants to evacuate the building (such as the rest station 
low in WTC 1 [somewhere between floors 12 and 20]). It remains unclear whether all of the 
mobility-impaired occupants and the helpers were able to successfully evacuate on September 11, 
2001. No evidence of a similar rest station in WTC 2 was found. 

While the mobility status of every decedent known to be below the impact region (107) in WTC 1 
could not be determined, it does not appear that mobility-impaired individuals were significantly 
overrepresented among the decedent population. 

Mobility-impaired occupants were not universally accounted for by existing evacuation 
procedures, as some were left by colleagues (later assisted by strangers), while others chose not to 
identify their mobility impairment to any colleagues. 



NIST NCSTAR 1-7, WTC Investigation 1 69 



Chapter 11 



This page intentionally left blank. 



1 70 NIST NCSTAR 1-7, WTC Investigation 



Chapter 12 

References 



1912. The Equitable Building Fire. Quarterly of the National Fire Protection Association 5, no. 4:408- 
445. 

1977. Triangle Shirtwaist Toll Of 145 Is Still Largest for U.S. Industrial Fires. Fire Engineering 130, no. 
8:55-58. 

Adler, J. 2002. Five Who Survived. Newsweek, September 9. 

ASME 2000. Safety Code for Elevators and Escalators ASME 17.1, American Society of Mechanical 
Engineers, New York. 

Blalock, H. 1972. Social Statistics. New York: McGraw Hill. 

BOMA 2001. Experience Exchange Report, Building Owners and Managers Association, New York. 

Bouchard, J. 1982. NFPA Summary Report on the Holiday Inn Fire, Kerney, Nebraska. NFPA Journal 
76, no. 6:20-24. 

Breasted, Mary. 1977. 100,000 Leave New York Offices As Bomb Threats Disrupt City; Blast Kills One 
and Hurt Seven. New York Times. 

Bukowski, R.W. and Kuligowski, E.D. 2003. The Basis for Egress Provision in U.S. Building Codes, 
Proceedings of InterFlam 2004, Interscience Communications, London. 

Bukowski, R.W. and Moore, W. D. 2003. Fire Alarm Signaling Systems. 3rd (Third) Edition. National 
Fire Protection Association, Quincy, MA. 

Bukowski, R. W. and W. D. Moore. 2003. Fire Alarm Signaling Systems. Quincy, MA: National Fire 
Protection Association. 

Cochrane, W. G. 1977. Sampling Techniques. New York: John Wiley & Sons. 

Couchon, D. 2001. For Many on Sept. 11, Survival Was No Accident . USA Today, December 19. 

Demers, D. 1980. Familiar Problems Cause 10 Deaths in Hotel Fire. NFPA Journal 74, no. 1:52-56. 

DiMarco, D. 2004. Tower Stories, The Autobiography of September 11, 2001, Revolution Publishing, 
New York. 

Fahy, R. F. 2001. Verifying the Predictive Capability of EXIT89. Paper presented at 2nd International 
Symposium on Human Behaviour in Fire. 



NIST NCSTAR 1-7, WTC Investigation 171 



Chapter 12 

Fahy, R. F. 1999. User's Manual, EXIT89 v 1.01, An Evacuation Model for High-Rise Buildings. 
Quincy, Ma, National Fire Protection Association. 

Fahy, R. F. and G. Proulx. 2003. Database of First-Person Accounts from Survivors of the World Trade 
Center Evacuation on September 11, 2001 conducted as part of report NIST NCSTAR 7-A. 

Fasullo, E. 1995. The Port Authority of New York and New Jersey, New York, personal communication 
to R. Visconti, New York City Dept. of Buildings. January 27. WTCI-159-P 

Fisher, R. P., K. H. Brennan, and M. R. McCauley. 2002. The Cognitive Interview Method to Enhance 
Eyewitness Recall. In Memory and Suggestibility in the Forensic Interview, edited by Eisen, M. L., J. 
A. Quas, and G. S. Goodman (Mahweh, New Jersey: Lawrence Erlbaum). 

Fisher, R. P., K. L. Falkner, M. Trevisan, and M. R. McCauley. 2000. Adapting the cognitive interview to 
enhance long-term (35 years) recall of physical activities. Journal of Applied Psychology 85, no. 
2:180-189. 

Fruin. J. J. 1987. Pedestrian Planning and Design, Revised Edition. Elevator World, Incorporated, 
Mobile, AL. 

Galea, E. R., P. Lawrence, S. J. Blake, S. Gwynne, and Westeng H 2004. A Preliminary Investigation of 
the Evacuation of the WTC North Tower Using Computer Simulation. 167-180. London, England, 
Interscience Communications Limited. 3rd International Symposium on Human Behaviour in Fire. 

Geiselman, R. E., R. P. Fisher, D. Mackinnon, and H. Holland. 1986. Enhancement of eyewitness 
memory with the cognitive interview. American Journal of Psychology 99:385-401. 

Gwynne, S., E.R. Galea, P. J. Lawrence, M. Owen, and L. Filippidis. 1999. A Review of the 

Methodologies used in the Computer Simulation of Evacuation from the Built Environment, Building 
and Environment, 34, 741-749. 

Gwynne, S., E. R. Galea, P. J. Lawrence, and L. Filippidis. 2001. Modeling Occupant Interaction with 
Fire Conditions Using the buildingEXODUS Model. Fire Safety Journal, 36, 327-357. 

Gwynne, S., E. R. Galea, P. J. Lawrence, M. Owen, and L. Filippidis. 1998. A Systematic Comparison of 
Model Prediction Produced by the buildingEXODUS Model and the Tsukuba Pavilion Evacuation 
Data, Applied Fire Science, Vol. 7, No. 3, 235-266. 

Hayne, W. G. 1945. Report of the Collision of U.S. Army Air Force Bomber with the Empire State 
Building at 78th and 79th Floors. 1-15. New York, New York Board of Fire Underwriters. 

IES 2000. Simulex Technical Reference; Evacuation Modeling Software. Integrated Environmental 
Solutions, Inc. 

IES 2001. Simulex User Manual; Evacuation Modeling Software. Integrated Environmental Solutions, 
Inc. 

Isner, M. S. 1990. Five Die in High-Rise Office Building Fire. NFPA Journal 84, no. 4:50-59. 

1 72 NIST NCSTAR 1 - 7, WTC Investigation 



References 



Isner, M. S. and T. J. Klem. World Trade Center Explosion and Fire, New York, New York, February 26, 
1993. Fire Investigation Report. 1993a. Quincy, MA, National Fire Protection Association. 

Isner, M. S. and T. J. Klem. 1993b. Explosion and Fire Disrupt World Trade Center. NFPA Journal:9l- 
104. 

Ivins, Molly. 1977. Most Stay Calm As Evacuations Halt Workday. New York Times. 

Jones, W. W., G. P. Forney, R. D. Peacock, and P. A. Reneke. Technical Reference for CFAST: An 
Engineering Tool for Estimating Fire and Smoke Transport. Tech. Note 1431. 2000. Natl. Inst. 
Stand. Technol. 

Keating, J. and E. Loftus. Post Fire Interviews. 1984. Development and Field Validation of the 
Behavioral Sequence Interview Technique. NBSIR 84-477. 1984. Nat. Bur. Stand. (U.S.). 

Klote, J. H. 1993. A Method of Calculation of Elevator Evacuation Time. Journal of Fire Protection 
Engineering 5, no. 3:83-95. 

Levy, M. P. 1965. The Port of New York Authority, New York, personal communication to M. 
Yamasaki, Minora Yamasaki & Associates, New York. September 29. 

Memorandum from John M. Kyle (PONYA) to Guy F. Tozzoli (PONY A), May 9, 1966. "World Trade 
Center - New York City Building Code - Proposed Revision." 

Murphy, Dean E. 2002. September 11: An Oral History. New York: Doubleday. 

New York 1973. Fire Safety in Office Buildings, New York City - Local Law 5, The City of New York. 
January 18. 

New York 1984. Fire Safety in Buildings, Including Hotels, New York City - Local Law 16, The City of 
New York. . 

New York 1995. The Report of the World Trade Center Review Committee. The City of New York. 

Nillson, L. 2000. Remembering actions and words. In The Oxford handbook of Memory, edited by 
Tulving, E. and F. I. M. Craik (New York: Oxford University Press). 

NIST 2004a. NIST WTC Face-to-Face Interview Data Set. 2004. Gaithersburg, MD, Natl. Inst. Stand. 
Technol. 

NIST 2004b. NIST WTC Emergency Responder Interview Data Set. 2004. Gaithersburg, MD, Natl. Inst. 
Stand. Technol. 

NYCBC 1968. Building Code - Local Law No. 76 of the City of New York, New York. 

NYCBC 2001. Building Code of the City of New York, 2001 Edition, Gould Publications, Binghamton, 

NY. 



NIST NCSTAR 1-7, WTC Investigation 1 73 



Chapter 12 

PANYNJ Undated. "WTC - CHANGES TO EVACUATION SYSTEMS AFTER 1993." The Port 
Authority of New York and New Jersey. WTC-1 15-P 

PANYNJ 1995. The Port Authority of NY & NJ World Trade Center Fire Safety Plan (Towers One and 
Two), The Port Authority of New York and New Jersey, New York. WTCI-13-NYC. 

PANYNJ 1996. WTC Fire Safety 1996. Video. WTCI-463-P 

PANYNJ 2001a. Port Authority of New York and New Jersey Radio Communications Transcripts from 
September 11,2001. 

PANYNJ 2001b World Trade Center Emergency Procedures Manual 2001 - Confidential. The Port 
Authority of New York and New Jersey. 

PANYNJ 2001c. World Trade Center Space Book Plan, Building 1 and Building 2, Port Authority Real 
Estate Department, The Port Authority of New York and New Jersey, New York. 

Powers, W. R. 1975. One World Trade Center Fire New York, N.Y., February 13, 1975. 1-15. Boston, 
MA, National Fire Protection Association. 

Powers, W. R. 1970. One New York Plaza Fire, New York, N.Y., August 5,1970. 1-15.. New York, New 
York Board of Fire Underwriters. 

Predtechenskii, V. M. and A. I. Milinskii. 1978. Planning for Foot Traffic in Buildings. New Delhi: 
Amerind Publishing Co. Pvt. Ltd. 

Schank, R. C. and R. Abelson. 1977. Scripts, Plans, Goals, and Understanding. Nillsdale, NJ: Erlbaum. 

Sharry, J. A. 1974. South America Burning. Fire Journal, 1974, 23-33. 

Solomon, J., 1968. Emory Roth and Sons, New York, personal communication to M. Levy, The Port of 
New York Authority, New York. January 25. WTCI-477-P 

Solomon, J., 1975. Emory Roth and Sons, New York, personal communication to M. Levy, The Port of 
New York Authority, New York. February 18. 

Templer, J. 1992. Staircase: History and Theories. MIT Press , Cambridge, MA. 

Willey, A. E. 1972. High Rise Building Fire, Sao Paulo, Brazil. Fire Journal, 1972, 7-108. 

Williams, F. D. 1990. SLAM: The Influence of S.L.A. Marwill on the United States Army. In TRADOC, 
(Ft. Monroe, VA: U. S. Army). 



1 74 NIST NCSTAR 1-7, WTC Investigation 



Appendix A 

Telephone Interview Script 



I would like to start by getting some background information. What year did you first start working at the 
World Trade Center? RANGE: 1975 - 2001 

$E 1975 2001 

DK9998 

RF 9999 

«YRWRK» 



On September 1 1th, 2001, were you in any of the following positions with the World Trade Center? 

Port Authority Staff 1 

Fire Safety Staff 2 

Floor Warden or Searcher 3 

Maintenance or Security Staff 4 

NONE OF THESE X 

DK8 

RF 9 

«ROLES_01» 

«ROLES_02» 

«ROLES_03» 

«ROLES_04» 

During the year from September 1 1th, 2000 to September 1 1th, 2001, how many fire drills did you take 
part in at the World Trade Center? 

$E0 99 

NONE 00 => SWLOC 

DK98 => SWLOC 

RF 99 => SWLOC 

«FIRED» 



During these drills, were you ever instructed about the location of the emergency stairwell nearest to your 
office? 

YES 1 



NIST NCSTAR 1-7, WTC Investigation 1 75 



Appendix A 

NO 2 => SWLOC 

DK 8 => SWLOC 

RF 9 => SWLOC 
«DEXIT» 

How many emergency stairwells were you shown? 

ONE 1 

TWO 2 

THREE 3 => LVFSW 

OTHER, SPECIFY 7 O 

DK8 

RF 9 

«HMEXT» 

«0 HMEXT» 



Before September 1 1th, had you learned in other ways about the locations of the three emergency 
stairwells? 

YES 1 

NO 2 

DK 8 

RF 9 

«SWLOC» 



SKIP IF NO FIRE DRILLS 
=> USESW 

Else=>+1 

ifFIRED=00,98-99 

«SOUTl» 

During any of the fire drills, did you leave your floor using one of the stairwells? 

YES 1 

NO 2 => USESW 

DK 8 => USESW 

RF 9 => USESW 

«LVFSW» 

1 76 NIST NCSTAR 1 - 7, WTC Investigation 



Telephone Interview Script 



Which stairwells did you use? 

STAIRWELL A 1 

STAIRWELL B 2 

STAIRWELL C 3 

OTHER, SPECIFY 7 O 

DK8 

RF 9 

«WHSW1_01» 

«WHSW1_02» 

«WHSW1_03» 

«WHSW1_04» 

«0_WHSW1» 

Which side of the building was the stairwell located on? 

=>+l 

if NOT WHS W 1=8 

NORTH 1 

SOUTH 2 

EAST 3 

WEST 4 

OTHER, SPECIFY 7 O 

DK8 

RF 9 

«WHSL1» 

«0_WHSL1» 

Had you ever used any of the emergency stairwells prior to September 1 1th? 

=> DHELP 

ifLVFSW=l 

YES 1 

NO 2 => DHELP 

DK 8 => DHELP 

RF 9 => DHELP 

«USESW» 

NIST NCSTAR 1-7, WTC Investigation 1 77 



Appendix A 



SKIP FOR NO DRILLS AND NO USE OF STAIRWELLS 

=> AEVOF 

Else=>+1 

if FIRED=00,98,99 AND USESW>1 

«SOUT2» 

Which stairwell did you use? 

STAIRWELL A 1 

STAIRWELL B 2 

STAIRWELL C 3 

OTHER, SPECIFY 7 O 

DK8 

RF 9 

«WHSW2_01» 

«WHSW2_02» 

«WHSW2_03» 

«WHSW2_04» 

«0_WHSW2» 

SKIP IF NO FIRE DRILLS 
=> AEVOF 

Else=>+1 

ifFIRED=00,98-99 

«SOUT3» 

When you were evacuating on September 1 1th, how helpful was your experience during these drills? 

=>+l 

ifFIRED=00 

Very Helpful 1 

Somewhat Helpful 2 

Somewhat Unhelpful 3 

Very Unhelpful 4 

DK8 

RF 9 

178 NIST NCSTAR 1 - 7, WTC Investigation 



Telephone Interview Script 



«DHELP» 

Prior to September 1 1th, were you aware of the evacuation procedures for your floor? 

YES 1 

NO 2 => FLWAR 

DK 8 => FLWAR 

RF 9 => FLWAR 

«AEVOF» 

Prior to September 1 1th, what was the evacuation procedure you were told to follow? 

LEAVE BUILDING IMMEDIATELY 1 

GO TO ELEVATOR LOBBY 2 

GO TO FLOORS UP OR DOWN 3 

GO TO ROOF 4 

STAY WHERE YOU ARE 5 

OTHER, SPECIFY 7 O 

DK8 

RF 9 

«EVACP» 

«0_EVACP» 

Did you know that there was a Floor Warden for your floor? 

=>+l 

ifROLES=l-4 

YES 1 

NO 2 

DK 8 

RF 9 

«FLWAR» 



The next questions ask about 3 different time periods. The first series of questions asks about when you 
first became aware that something had happened at the World Trade Center. This is a period of just a few 
seconds. The next series of questions asks about the time from when you first became aware that 
something had happened, to the time you first entered a stairwell or elevator to exit the building. The 
third series of questions asks about what happened during your evacuation, meaning the time from when 



NIST NCSTAR 1-7, WTC Investigation 1 79 



Appendix A 



you first entered a stairwell or elevator until you exited the tower. At the end of the interview, I will ask 
you if there is anything else about your experience on September 1 1th that you would like to contribute. 

CONTINUE 1 D 

«IFAWA» 



Now thinking back to the morning of September 1 1th, how did you first become aware that something 
had happened at the World Trade Center? 

$E 1 9 

HEARD SOMETHING (BOOM, CRASH, EXPLOSION, BLAST, ROAR, RUMBLING, ALARM) 
01 

SAW SMOKE OR FLAMES 02 

SAW DEAD BODIES 03 

SAW A PLANE 04 

SAW DEBRIS 05 

FELT SOMETHING (BUILDING MOVING, IMPACT, SHAKING, SWAYING, ROCKING, JOLT, 

EARTHQUAKE) 06 

FELL DOWN/FELL OFF CHAIR 07 

WARNED BY SOMEONE AROUND ME 08 

CONTACTED VIA PHONE 09 

CONTACTED VIA EMAIL 1 

PUBLIC ADDRESS SYSTEM 11 

NEWS MEDIA (TELEVISION, RADIO) 12 

OFFICE FURNITURE OR FIXTURES FALLING 1 3 

FURNITURE OR OTHER ITEMS FALLING OVER/DOWN 1 4 

OTHER, SPECIFY 97 O 

DK98 

RF 99 

«FAWAR» 

«0_FAWAR» 

What were you doing when you first became aware that something had happened to the World Trade 
Center? PROBE: Anything else? 

$E 1 9 

WORKING INDEPENDENTLY 01 

IN MEETING 02 

ON PHONE 03 



180 NIST NCSTAR 1-7, WTC Investigation 



Telephone Interview Script 



CHECKING/WRITING EMAIL 

WAITING FOR ELEVATOR 05 

RIDING IN ELEVATOR 06 

CHATTING WITH COWORKERS 

EATING/HAVING COFFEE 08 

ENTERING BUILDING 09 

OTHER, SPECIFY 97 O 

DK98 X 

RF 99 X 

«ACTV1_01» 

«ACTV1_02» 

«ACTV1_03» 

«ACTV1_04» 

«ACTV1_05» 

«ACTV1_06» 

«ACTV1_07» 

«ACTV1_08» 

«ACTV1_09» 

«ACTV1_10» 

«0 ACTV1» 



04 



07 



At the moment when you first became aware that something had happened at the World Trade Center, did 
you notice any of the following? FOLLOW UP: Was that in your immediate area or outside the Tower? 




Did Not Notice 


Noticed in Immediate Area 


Noticed Outside the 
Tower 


Smoke 








Fire or Flames 








Fireballs 








Collapsed walls 








Jet Fuel 








Severely or fatally injured people 








Sprinklers going on 








A fire alarm sounding 








Power outage or flickering lights 








Fallen ceiling tiles 








Extreme heat 









NISTNCSTAR 1-7, WTC Investigation 



181 



Appendix A 

«NOT01_01» 
«NOT01_02» 

TIME PERIOD: 1 

Were there any disaster related events going on around you at this time? 

=> WHTW2 

ifOR[NOT01-NOTll]=2-3 

YES 1 

NO 2 => WHTW2 

DK 8 => WHTW2 

RF 9 => WHTW2 

«OEVEN» 

TIME PERIOD: 1 

What was going on? 

ENTER RESPONSE 1 O 

DK8 

RF 9 

«GOING» 

«0_GOING» 

TIME PERIOD: 1 

Were you still in<WHTOW>at this time? IF YES, SELECT APPROPRIATE CHOICE IF NO, ASK 
WHICH TOWER THEY WERE IN 

TOWER 1 1 

TOWER 2 2 

DK8 

RF 9 

«WHTW2» 

TIME PERIOD: 1 

And were you still on the<WHFLO>floor at this time? RANGE: 1st - 1 10th FLOOR IF YES, 
SELECT/ENTER FLOOR IF NO, ASK WHICH FLOOR THEY WERE ON AND SELECT/ENTER IT 

$E 1 110 

BASEMENT 990 

1 82 NIST NCSTAR 1 - 7, WTC Investigation 



Telephone Interview Script 



CONCOURSE/LOBBY 99 1 

PLAZA992 

IN ELEVATOR 993 

OTHER, SPECIFY 997 O 

DK998 

RF 999 

«WHFL2» 

«0_WHFL2» 

TIME PERIOD: 1 

At the moment when you first became aware that something had happened to the World Trade Center, 
approximately how many people were with you? RANGE: - 999 PEOPLE WE WANT THE 
NUMBER OF PEOPLE THAT WERE IN THE SAME LOCATION AS THE RESPONDENT. (IN 
THEIR LINE OF SIGHT) 

$E 999 

NONE 00 =>YOUIN 

DK 98 => YOUIN 

RF 99 => YOUIN 

«PEOPl» 

TIME PERIOD: 1 

Were any of these people injured at that time as a result of the event? 

YES 1 

NO 2 

DK 8 

RF 9 

«PEOESr» 

TIME PERIOD: 1 

Were you injured at that time, as a result of the event? 

YES 1 

NO 2 => ORISK 

DK 8 => ORISK 

RF 9 => ORISK 

«YOUIN» 



NIST NCSTAR 1-7, WTC Investigation 1 83 



Appendix A 

TIME PERIOD: 1 

Would you say your injury was a . . . 

An injury that did not impact your ability to evacuate, 1 

An injury that did impact your ability to evacuate but was not life threatening, or 2 

A life threatening injury 3 

OTHER, SPECIFY 7 O 

DK8 

RF 9 

«NATIN» 

«0_NATIN» 

TIME PERIOD: 1 

Still thinking about the moment when you first became aware that something had happened at the World 
Trade Center, did you believe that other people were in danger of being killed? 

YES 1 

NO 2 

DK 8 

RF 9 

«ORISK» 

TIME PERIOD: 1 

Did you believe you were in danger of being killed? 

YES 1 

NO 2 

DK 8 

RF 9 

«YRISK» 

TIME PERIOD: 2 

Now please think about the time period between when you first became aware that something had 
happened and when you first entered a stairwell or elevator to leave the tower. During this entire time 
period, were you given any additional information about what was going on? AFTER BECOMING 
AWARE OF THE EVENT, BUT BEFORE EVACUATION 

YES 1 

NO 2 => SEEKI 

DK 8 => SEEKI 

1 84 NIST NCSTAR 1 - 7, WTC Investigation 



Telephone Interview Script 



RF 9 => SEEKI 

«GETIN» 

TIME PERIOD: 2 

Who gave you this information? PROBE: Anyone else? 

MANAGER/SUPERVISOR 1 

COWORKER INSIDE BUILDING 2 

FAMILY/FRIEND OUTSIDE BUILDING 3 

POLICE/FIREFIGHTER 4 

FLOOR WARDEN 5 

MEDIA PERSON (TV/RADIO) 6 

OTHER, SPECIFY 7 O 

DK8 X 

RF 9 X 

«WHINF_01» 

«WHINF_02» 

«WHINF_03» 

«WHINF_04» 

«WHINF_05» 

«WHINF_06» 

«WHINF_07» 

«0_WHINF» 

TIME PERIOD: 2 

What information did you get? PROBE: Any other information? 

INFORMATION ABOUT WHAT HAD HAPPENED 1 

INSTRUCTIONS TO LEAVE 2 

INSTRUCTIONS TO STAY 3 

OTHER, SPECIFY 7 O 

DK8 X 

RF 9 X 

«WHATI_01» 

«WHATI_02» 

«WHATI_03» 

«WHATI_04» 

NIST NCSTAR 1-7, WTC Investigation 1 85 



Appendix A 

«0_WHATI» 

TIME PERIOD: 2 

How did you get this information? PROBE: Any other way? 

FACE TO FACE 1 

TELEPHONE 2 

EMAIL/BLACKBERRY 3 

PA ANNOUNCMENT 4 

TV/RADIO 5 

OTHER, SPECIFY 7 O 

DK8 X 

RF 9 X 

«HOWGT_01» 

«HOWGT_02» 

«HOWGT_03» 

«HOWGT_04» 

«HOWGT_05» 

«HOWGT_06» 

«0_HOWGT» 

TIME PERIOD: 2 

And during this same time period, did you try to get additional information about what was going on? 
AFTER BECOMING AWARE OF THE EVENT, BUT BEFORE EVACUATION 

YES 1 

NO 2 => ORIS2 

TRIED, BUT WAS UNABLE TO GET INFORMATION 3 => ORIS2 

DK 8 => ORIS2 

RF 9 => ORIS2 

«SEEKI» 

TIME PERIOD: 2 

Who did you go to for this information? PROBE: Anyone else? 

MANAGER/SUPERVISOR 1 

COWORKER INSIDE BUILDING 2 

FAMILY/FRIEND OUTSIDE BUILDING 3 

186 NIST NCSTAR 1 - 7, WTC Investigation 



Telephone Interview Script 



POLICE/FIREFIGHTER 4 

FLOOR WARDEN 5 

MEDIA PERSON (TV/RADIO) 6 

OTHER, SPECIFY 7 O 

DK8 X 

RF 9 X 

«GOINF_01» 

«GOINF_02» 

«GOINF_03» 

«GOINF_04» 

«GOINF_05» 

«GOINF_06» 

«GOINF_07» 

«0_GOINF» 

TIME PERIOD: 2 

What type of information did you try to find? PROBE: Anything else? 

INFORMATION ABOUT WHAT HAD HAPPENED 1 

INSTRUCTIONS TO LEAVE 2 

INSTRUCTIONS TO STAY 3 

OTHER, SPECIFY 7 O 

DK8 X 

RF 9 X 

«WHAI2_01» 

«WHAI2_02» 

«WHAI2_03» 

«WHAI2_04» 

«0_WHAI2» 

TIME PERIOD: 2 

How did you get this information? PROBE: Any other way? 

FACE TO FACE 1 

TELEPHONE 2 

EMAIL/BLACKBERRY 3 

PA ANNOUNCMENT 4 

NIST NCSTAR 1-7, WTC Investigation 1 87 



Appendix A 

TV/RADIO 5 

OTHER, SPECIFY 7 O 

DK8 X 

RF 9 X 

«HOWG2_01» 

«HOWG2_02» 

«HOWG2_03» 

«HOWG2_04» 

«HOWG2_05» 

«HOWG2_06» 

«0_HOWG2» 

TIME PERIOD: 2 

And during the time between when you first became aware that something had happened at the World 
Trade Center and when you first entered the stairwell or elevator to leave the tower, did you believe that 
other people were in danger of being killed? AFTER BECOMING AWARE OF THE EVENT, BUT 
BEFORE EVACUATION 

=> YRIS2 

ifORISK=l 

YES 1 

NO 2 

DK 8 

RF 9 

«ORIS2» 

TIME PERIOD: 2 

During that time period, did you believe you were in danger of being killed? 

=> PEODO 

ifYRISK=l 

YES 1 

NO 2 

DK 8 

RF 9 

«YRIS2» 

TIME PERIOD: 2 

1 88 NIST NCSTAR 1-7, WTC Investigation 



Telephone Interview Script 



During this time period, what were the people around you doing? PROBE: Were they doing anything 
else? AFTER BECOMING AWARE OF THE EVENT, BUT BEFORE EVACUATION 

$E0 10 

NOONE AROUND/WAS ALONE 00 X 

TALKING TO OTHERS 1 

GATHERING PERSONAL/WORK ITEMS 02 

SEARCHING FOR OTHERS 03 

CALLING OTHERS 04 

FIGHTING FIRE/SMOKE 05 

LOCKING UP 06 

WORKING 07 

EVACUATING THE TOWER 08 

CRYING, RUNNING AROUND, IN SHOCK 09 

HELPING OTHERS 10 

OTHER, SPECIFY 97 O 

DK98 X 

RF 99 X 

«PEODO_01» 

«PEODO_02» 

«PEODO_03» 

«PEODO_04» 

«PEODO_05» 

«PEODO_06» 

«PEODO_07» 

«PEODO_08» 

«PEODO_09» 

«PEODO_10» 

«0_PEODO» 

TIME PERIOD: 2 

Did the people around you start evacuating before you did? 

=> DOBEF 

ifPEODO=08 

YES 1 

NO 2 

NIST NCSTAR 1-7, WTC Investigation 1 89 



Appendix A 

DK 8 
RF 9 
«EVACB» 

TIME PERIOD: 2 

Did you do any of the following before starting your evacuation? 

$E 1 9 

Talk to another person face to face 1 

Gather personal items 02 

Telephone other people 03 

Continue working 04 

Save or transfer computer files 05 

Search for others 06 

Fight fire or smoke 07 

Move to another floor 08 

Help others 09 

Logging off/shutting down computer 10 

NONE OF THESE 11 X 

«DOBEF_01» 

«DOBEF_02» 

«DOBEF_03» 

«DOBEF_04» 

«DOBEF_05» 

«DOBEF_06» 

«DOBEF_07» 

«DOBEF_08» 

TIME PERIOD: 2 

Did you do anything else during this time? 

ENTER RESPONSE 1 O 

NO OTHER ACTIVITIES 

DK8 

RF 9 

«OACTI» 

«0 OACTI» 



190 



NIST NCSTAR 1-7, WTC Investigation 



Telephone Interview Script 



TIME PERIOD: 2 

Before you began your evacuation, was there anything you wanted to do, but couldn't? 

YES 1 

NO 2 => SEE01 

DK 8 => SEE01 

RF 9 => SEE01 

«WANTD» 

TIME PERIOD: 2 

What was that? PROBE: Anything else? 

$E 1 7 

GATHER WORK ITEMS 01 

GATHER PERSONAL BELONGINGS 02 

CALL FRIEND/FAMILY MEMBER 03 

FIND FRIEND/COWORKER 04 

HELP FRIEND/COWORKER 05 

LOCKUP 06 

EVACUATE IMMEDIATELY 07 

OTHER, SPECIFY 97 O 

DK98 X 

RF 99 X 

«WANAC_01» 

«WANAC_02» 

«WANAC_03» 

«WANAC_04» 

«WANAC_05» 

«WANAC_06» 

«WANAC_07» 

«WANAC_08» 

«0_WANAC» 

TIME PERIOD: 2 

Why couldn't you do that/those things? 

$E 1 9 

NIST NCSTAR 1-7, WTC Investigation 191 



Appendix A 



AFRAID 01 

LOCKED DOORS 02 

PHONE LINES DEAD 03 

INJURED 04 

EXIT BLOCKED 05 

TOO CROWDED 06 

TOLD TO STAY IN BUILDING 

TOLD TO LEAVE 08 

FATIGUE 09 

DISABLED 10 

SMOKE 1 1 

DAMAGE TO FLOOR 12 

WAS HELPING OTHERS 13 

OTHER, SPECIFY 97 O 

DK98 

RF 99 

«WHYNO_01» 

«WHYNO_02>: 

«WHYNO_03>: 

«WHYNO_04>: 

«WHYNO_05>: 

«WHYNO_06>: 

«WHYNO_07>: 

«WHYNO_08>: 

«WHYNO_09>: 

«WHYNO_10>: 

«WHYNO_ll>: 

«WHYNO_12>: 

«WHYNO_13>: 

«WHYNO_14>: 

«0 WHYNO» 



07 



192 



NIST NCSTAR 1-7, WTC Investigation 



Telephone Interview Script 





Still thinking about the time between when you first became aware that something had happened at the 
World Trade Center and when you entered the stairwell or elevator to leave the tower, did you notice any 
of the following? FOLLOW UP: Was that in your immediate area or outside the Tower? 




Did Not Notice 


Noticed in Immediate Area 


Noticed Outside the 
Tower 


Smoke 








Fire or Flames 








Fireballs 








Collapsed walls 








Jet Fuel 








Severely or fatally injured 
people 








Sprinklers going on 








A fire alarm sounding 








Power outage or flickering lights 








Fallen ceiling tiles 








Extreme heat 









«SEE01_01» 

«SEE01_02» 

TIME PERIOD: 2 

Were there any disaster related events going on around you at this time? 

=> EVACF 

ifOR[SEE01-SEEll]=2-3 

YES 1 

NO 2 => HELPY 

DK 8 => HELPY 

RF 9 => HELPY 

«ODISE» 

TIME PERIOD: 2 

What was going on? 

ENTER RESPONSE 1 O 

DK8 

RF 9 

«GOIN2» 

«0 GOIN2» 



NISTNCSTAR 1-7, WTC Investigation 



193 



Appendix A 



TIME PERIOD: 2 

Were you still on the<WHFL2>floor at this time? RANGE: 1st - 1 10th FLOOR IF YES, 
SELECT/ENTER FLOOR IF NO, ASK WHICH FLOOR THEY WERE ON AND SELECT/ENTER IT 

$E 1 110 

=>+l 

if (AND[SEE01-SEE11]=1) AND PEODO>0 AND PEODO<98 

BASEMENT 990 

CONCOURSE/LOBBY 991 

PLAZA 992 

ELEVATOR 993 

OTHER, SPECIFY 997 O 

DK998 

RF 999 

«EVACF» 

«0_EVACF» 

TIME PERIOD: 2 

Did anyone help you in any way before you started your evacuation? 

YES 1 

NO 2 => DECID 

DK 8 => DECID 

RF 9 => DECID 

«HELPY» 

TIME PERIOD: 2 

Who helped you? PROBE: Anyone else? WE WANT THEIR ROLE NOT THE NAME OF THE 
PERSON 

POLICE OFFICER/FIREFIGHTER 1 

COWORKER 2 

STRANGER 3 

FLOOR WARDEN 4 

MANAGER/SUPERVISOR 5 

OTHER, SPECIFY 7 O 

DK8 X 



1 94 NIST NCSTAR 1-7, WTC Investigation 



Telephone Interview Script 



RF 9 X 

«WHOHE_01» 

«WHOHE_02» 

«WHOHE_03» 

«WHOHE_04» 

«WHOHE_05» 

«WHOHE_06» 

«0_WHOHE» 

TIME PERIOD: 2 

What did they help you with? PROBE: Anything else? 

$E 1 7 

LOCATING OTHERS 01 

HELPING OTHERS 02 

FINDING EXITS 03 

TREATING YOUR INJURIES 04 

PROVIDED INFORMATION/INSTRUCTIONS 05 

GATHER BELONGINGS 06 

CALM DOWN/EMOTIONAL ASSISTANCE 07 

OTHER, SPECIFY 97 O 

DK98 X 

RF 99 X 

«WHATD_01» 

«WHATD_02» 

«WHATD_03» 

«WHATD_04» 

«WHATD_05» 

«WHATD_06» 

«WHATD_07» 

«WHATD_08» 

«0_WHATD» 

TIME PERIOD: 2 

What was the one thing that made you decide to evacuate? 

WAS TOLD TO EVACUATE 1 

NIST NCSTAR 1-7, WTC Investigation 1 95 



Appendix A 

FRIENDS CO-WORKERS EVACUATED 

AFRAID/FELT IN DANGER 3 

FIRE ALARM WAS GOING OFF 4 

SAW SMOKE 5 

SAW FIRE 6 

OTHER, SPECIFY 7 O 

DK8 

RF 9 

«DECID» 

«0 DECID» 



How many minutes had passed before you started to evacuate? IF NEEDED: How much time passed 
between when you first became aware that something had happened to the World Trade Center and when 
you entered the stairwell or elevator to leave the tower. THIS IS NOT TIME TO EVACUATE. 
PLEASE CLARIFY WITH RESPONDENT IF TIME APPEARS TOO LONG. RESPONDENT WAS 
IN<WHTW2> RANGE FOR TOWER 1: 1 - 103 MINUTES RANGE FOR TOWER 2: 1 - 75 
MINUTES 

$E 1 103 

DK 998 

RF 999 

«TIMEP» 



SKIP FOR TOWERS 
=> EVAC2 

Else=>+1 

ifWHTW2=2 

«SKIP1» 

Did you begin your evacuation. . . WE ARE INTERESTED IN WHAT THEY KNOW NOW. THEY 
MAY NOT HAVE KNOWN WHEN THEY WERE EVACUATING, BUT NOW THEY CAN TELL US 
WHEN IT WAS. 

Before the plane hit Tower 2 1 

After the plane hit Tower 2, but before the Tower 2 collaspe 2 

After the Tower 2 collaspe 3 

DK 8 

RF 9 

«EVAC1» 

1 96 NIST NCSTAR 1 - 7, WTC Investigation 



Telephone Interview Script 



SELECT1 

$S NS=2 C0=1 IN=EVAC1<=1 ;C0=2 IN=EVAC1<=2 ; 

Before the plane hit Tower 2 1 

After the plane hit Tower 2, but before the Tower 2 collaspe 2 

After the Tower 2 collaspe 3 

DK8 

RF 9 

«SEL1» 

SELECT2 

Before the plane hit Tower 2 1 

After the plane hit Tower 2, but before the Tower 2 collaspe 2 

After the Tower 2 collaspe 3 

DK8 

RF 9 

«SEL2» 

Did you begin your evacuation. . . 

=> EVCSO 

ifEVACl>0 

Before the plane hit Tower 2 1 

After the plane hit Tower 2 2 

DK8 

RF 9 

«EVAC2» 

SELECT4 

$S CO=l IN=EVAC2<=1 ; 

Before the plane hit Tower 2 1 

After the plane hit Tower 2 2 

DK8 

RF 9 

«SEL3» 



NIST NCSTAR 1-7, WTC Investigation 1 97 



Appendix A 

Was there anything that kept you from evacuating sooner? 

YES, RECORD RESPONSE 1 O 

NO 2 

DK8 

RF 9 

«EVCSO» 

«0_EVCSO» 

TIME PERIOD: 3 

When you began your evacuation, were you alone or with other people? PEOPLE THAT THEY KNOW, 
PEOPLE THAT THEY WERE TALKING WITH 

ALONE 1 

WITH OTHER PEOPLE 2 

DK8 

RF 9 

«ALONE» 

TIME PERIOD: 3 

Which stairwell did you use for your evacuation? 



STAIRWELL A 


1 




STAIRWELL B 


2 




STAIRWELL C 


3 




USED ELEVATOR 


4 


=> FOLA1 


OTHER, SPECIFY 


7 




DK8 X 






RF 9 X 






«STAIR_01» 






«STAIR_02» 






«STAIR_03» 






«STAIR_04» 






«STAIR_05» 






«0 STAIR» 







TIME PERIOD: 3 

Which side of the building was the stairwell located on? 

198 NIST NCSTAR 1 - 7, WTC Investigation 



Telephone Interview Script 



=> /WHYST 

if NOT STAIR=8,7 

NORTH 1 

SOUTH 2 

EAST 3 

WEST 4 

OTHER, SPECIFY 7 O 

DK8 

RF 9 

«WHISI» 

«0_WHISI» 

TIME PERIOD: 3 

Why did you choose that/those stairwell(s) for your evacuation? PROBE: Any other reason? 

CLOSEST ONE 1 

FOLLOWED OTHER PEOPLE TO IT 2 

OTHER EXITS WERE BLOCKED 3 

SAME AS I USED IN PREVIOUS EMERGENCY 4 

I WAS TOLD TO USE THIS STAIRWELL 5 

OTHER, SPECIFY 7 O 

DK8 X 

RF 9 X 

«WHYST_01» 

«WHYST_02» 

«WHYST_03» 

«WHYST_04» 

«WHYST_05» 

«WHYST_06» 

«0_WHYST» 

TIME PERIOD: 3 

At any time during your evacuation, did you leave that/those stairwell(s)? DO NOT INCLUDE PEOPLE 
WHO FOLLOWED THE PASSAGE WHERE THE STAIRWELLS START AND END. 

YES 1 

NO 2 => FOLA1 

NIST NCSTAR 1-7, WTC Investigation 1 99 



Appendix A 

DK8 =>F0LA1 

RF 9 => F0LA1 

«LEVST» 

TIME PERIOD: 3 

Which floor were you on when you left the stairwell? IF RESPONDENT UNSURE, SELECT 997 AND 
RECORD RANGE OF FLOORS EXAMPLE: 34-40 

$R1 110 

UNSURE, RECORD RESPONSE 997 O 

«FLLST» 

«0_FLLST» 

TIME PERIOD: 3 

Why did you leave the stairwell? PROBE: Any other reason? 

$E 1 9 

I GOT LOST 01 

WAS TOLD TO LEAVE STAIRWELL 02 

TO HELP SOMEONE 03 

TO GO BACK AND GET SOMETHING 04 

TOO CROWDED 05 

SMOKE IN STAIRWELL 06 

PATH OBSTRUCTED 07 

A LOCKED DOOR 08 

STAIRWELL LED TO A FLOOR 09 

OTHER, SPECIFY 97 O 

DK98 

RF 99 

«WHYLS_01» 

«WHYLS_02» 

«WHYLS_03» 

«WHYLS_04» 

«WHYLS_05» 

«WHYLS_06» 

«WHYLS_07» 

«WHYLS_08» 

200 NIST NCSTAR 1 - 7, WTC Investigation 



Telephone Interview Script 



«WHYLS_09» 
«WHYLS_10» 
«0_WHYLS» 

Screen [Template 3] -> FL0A5 

=>+l 

ifFLWAR>l 



Did any of the following help you evacuate while you were in the building? 




Yes No DK RF 


Instructions or assistance from your floor warden 










Instructions or assistance from Police or Firefighters 










Support and encouragement from others 










Exit signs 










Photo luminescent paint in stairwells 











«FOLAl» 

Screen [Template 3] -> EVCM7 

=>+l 

if NOT STAIR<4 



Did any of the following make your evacuation more difficult while you were in the building? 




Yes 


No 


DK 


RF 


Crowded stairwells 










Firefighters or Police moving up stairwell 










Disabled or injured people being taken down stairwell 










Locked doors 










Poor lighting 










Confusing or missing signs 










Lack of clear instructions 











«EVCM1» 



Screen [Template 3] -> EXP1 1 



NISTNCSTAR 1-7, WTC Investigation 



201 



Appendix A 





Please tell me if you noticed any of the following at any time during your evacuation. FOLLOW UP: 
Was that in your immediate area or outside the Tower? 




Did Not Notice 


Noticed in Immediate Area 


Noticed Outside the Tower 


Smoke 








Fire or Flames 








Fireballs 








Collapsed walls 








Jet Fuel 








Severely or fatally injured people 








Sprinklers going on 








A fire alarm sounding 








Power outage or flickering lights 








Fallen ceiling tiles 








Extreme heat 









«EXP01_01» 
«EXP01_02» 

TIME PERIOD: 3 

During your evacuation, did you turn back at any time? "TURN BACK" MEANS "GO BACK UP". 

YES 1 

NO 2 => EXITS 

DK 8 => EXITS 

RF 9 => EXITS 

«TURNB» 

TIME PERIOD: 3 

Why did you turn back? PROBE: Any other reason? 

$E 1 7 

I GOT LOST 01 

I WAS TOLD TO TURN BACK 02 

TO HELP SOMEONE 03 

TO GET SOMETHING 04 

IT WAS TOO CROWDED 05 

SMOKE IN THE STAIRWELL 06 



202 



NIST NCSTAR 1-7, WTC Investigation 



Telephone Interview Script 



MY PATH WAS OBSTRUCTED 07 

OTHER, SPECIFY 97 O 

DK98 X 

RF 99 X 

«WHYTB_01» 

«WHYTB_02» 

«WHYTB_03» 

«WHYTB_04» 

«WHYTB_05» 

«WHYTB_06» 

«WHYTB_07» 

«WHYTB_08» 

«0_WHYTB» 

TIME PERIOD: 3 

Did you exit the stairwell or elevator to the mezzanine or to the concourse? 

MEZZANINE 1 

CONCOURSE 2 

OTHER, SPECIFY 7 O 

DK8 

RF 9 

«EXITS» 

«0_EXITS» 

TIME PERIOD: 3 

How much time passed between the moment you first began your evacuation to when you exited the 
Tower? PLEASE CLARIFY WITH RESPONDENT IF TIME APPEARS TOO LONG. RESPONDENT 
WAS IN<WHTW2> RANGE FOR TOWER 1: 1 - 103 MINUTES RANGE FOR TOWER 2: 1 - 75 
MINUTES 

$E 1 103 

DK 998 

RF 999 

«TIMP2» 

SKIP FOR TOWERS 

=>+2 

NIST NCSTAR 1-7, WTC Investigation 203 



Appendix A 

Else=>+1 

ifWHTW2=2 

«SKIP2» 

TIME PERIOD: 3 

Did you exit the tower. . . 

Eliminate -> 2 

According to NOT SEL1-SEL2 

Before the plane hit Tower 2 1 => GETOU 

After the plane hit Tower 2 but before the Tower 2 collapse, or 2 => GETOU 

After the Tower 2 collapse 3 => GETOU 

DK 8 => GETOU 

RF 9 => GETOU 

«EXIT1» 

TIME PERIOD: 3 

Did you exit the tower. . . 

Eliminate -> 1 

According to NOT SEL3 

Before the plane hit Tower 2, or 1 

After the plane hit Tower 2 2 

DK8 

RF 9 

«EXIT2» 



Please remember that this study is intended as a fact finding mission and not a fault finding mission. It is 
crucial that we determine why some people were successful in their evacuation while others were not. 
Was there anyone on your floor that was not successful in their evacuation? 

YES 1 



NO 2 


=> PHYSI 


DK 8 


=> PHYSI 


RF 9 


=> PHYSI 


«GETOU» 





Why didn't they make it out? PROBE: Any other reason? 



204 NIST NCSTAR 1-7, WTC Investigation 



Telephone Interview Script 



$E 1 8 

WAS INJURED 01 

WAS DISABLED 02 

REFUSED TO LEAVE 03 

DID NOT THINK IT WAS SERIOUS 04 

STAYED BACK TO HELP SOMEONE 05 

WAS TOLD TO STAY 06 

STRUCTURAL DAMAGE 07 

SMOKE OR FIRE 08 

OTHER, SPECIFY 97 O 

DK98 X 

RF 99 X 

«WHYNG_01» 

«WHYNG_02» 

«WHYNG_03» 

«WHYNG_04» 

«WHYNG_05» 

«WHYNG_06» 

«WHYNG_07» 

«WHYNG_08» 

«WHYNG_09» 

«0 WHYNG» 



On September 1 1th, 2001, did you have any physical problems that made it more difficult for you to leave 
the tower? Please do not include injuries caused by the incident or evacuation. 



YES 1 




NO 2 


=>AGE 


DK 8 


=>AGE 


RF 9 


=>AGE 


«PHYSI» 





What type of physical problem? PROBE: Anything else? 

$E 1 9 

BLIND/PARTIALLY BLIND 01 

DEAF 02 

NIST NCSTAR 1-7, WTC Investigation 205 



Appendix A 



IN WHEELCHAIR 03 

NEED WALKING ASSISTANCE 

OBESITY 05 

HEART CONDITION 06 

PREGNANT 07 

ASTHMA 08 

ELDERLY 09 

OTHER, SPECIFY 97 O 

DK98 X 

RF 99 X 

«LIMIT_01» 

«LIMIT_02» 

«LIMIT_03» 

«LIMIT_04» 

«LIMIT_05» 

«LIMIT_06» 

«LIMIT_07» 

«LIMIT_08» 

«LIMIT_09» 

«LIMIT_10» 

«0 LIMIT» 



04 



What is your age? RANGE: 1 - 98 YEARS 
$E 1 99 
RF 99 
«AGE» 

READ ONLY IF YOU CAN'T TELL. What is your gender? 

MALE 1 

FEMALE 2 

RF 9 

«GEND» 

What language do you speak best? 
ENGLISH 1 



206 



NIST NCSTAR 1-7, WTC Investigation 



Telephone Interview Script 



SPANISH 2 

OTHER, SPECIFY 7 O 

DK8 

RF 9 

«PLANG» 

«0_PLANG» 

Were you working in Tower 1 or Tower 2 during the 1993 bombing? 

=>SAY11 

ifYRWRK>1993 

YES 1 

NO 2 => CONCR 

DK 8 => CONCR 

RF 9 => CONCR 

«WBOMB» 

During the 1 993 bombing, did you evacuate immediately or wait to evacuate? 

EVACUATE IMMEDIATELY 1 

WAIT TO EVACUATE 2 

DK 8 => +2 

RF 9 => +2 

«EVBOM» 

At the time of the 1993 bombing, did you feel you that your decision to<EVBOM>was the right decision? 

YES 1 

NO 2 

DK 8 

RF 9 

«DEC93» 

After the 1993 bombing how concerned were you that terrorists would attack the World Trade Center? 
Were you... 

Extremely Concerned 1 

Very Concerned 2 

Moderately Concerned 3 

NIST NCSTAR 1-7, WTC Investigation 207 



Appendix A 

Slightly Concerned 4 

Not at all Concerned 5 

DK8 

RF 9 

«CONCR» 

Is there anything else you would like to say about your experience on September 1 1th? 

YES, RECORD RESPONSE 1 O 

NO 2 

DK 8 

RF 9 

«SAY11» 

«0_SAY11» 

IMPACT FLOOR FLAG 

=> * 

if IF(((WHTW2=1 AND WHFL2>91 AND WHFL2<99) OR (WHTW2=2 AND WHFL2>77 AND 
WHFL2<111)),1,0) 

IMPACT FLOOR FLAG 1 

«FFLAG» 

163: LFLAG 

Single 
min = 1 max =1 1=1 
2003/09/18 15:21 
LOCATION FLAG 
=> * 

if IF((WHFL2>990 AND WHFLK994), 1 ,0) 
LOCATION FLAG 1 
«LFLAG» 

EVENT FLAG 
=> * 

if IF(((AND[NOT02-NOT06]=2-3) OR (AND[SEE02-SEE06]=2-3) OR (AND[EXP02-EXP06]=2- 
3)), 1,0) 

208 NIST NCSTAR 1 - 7, WTC Investigation 



Telephone Interview Script 



EVENT FLAG 1 
«EFLAG» 

DISABILITY FLAG 
=> * 

ifIF((PHYSI=l),l,0) 
DISABILITY FLAG 1 
«DFLAG» 

ROLE FLAG 
=> * 

ifIF((ROLES=l-4),l,0) 
ROLE FLAG 1 
«RFLAG» 

We may be interested in learning more about your experience on September 1 1th. Would it be okay if we 
follow up with you sometime in the future to get more detailed information on your evacuation 
experience? 

=>+l 

ifFFLAG+LFLAG+EFLAG+DFLAG+RFLAG==0 

YES 1 

NO 2 

«FOLUP» 

PRESS ENTER TO CONTINUE 

Those are all the questions we have. The valuable information you provided will help designers and 
engineers improve building safety, and help emergency planners improve building evacuation procedures. 
Thank you so much for taking the time to talk with me, and have a good day/evening. Good-bye. 

END OF SURVEY 1 D => /INT99 

«THANK» 



NISTNCSTAR 1-7, WTC Investigation 209 



Appendix A 



This page intentionally left blank. 



210 NIST NCSTAR 1-7, WTC Investigation 



Appendix B 

Qualitative Analysis Coding 



The qualitative data from the face-to-face interviews was used to enrich the statistically-based telephone 
interview data by providing detailed descriptions of the experiences of occupants of the two towers on 
September 1 1 , 200 1 . These data identified unknown information, evaluated technical hypotheses, and 
explored conscious and subconscious motivations for occupant behaviors, while allowing for 
comparisons to the telephone interview data. This appendix provides an annotated listing of the coding 
used by National Institute Standards and Technology (NIST) to analyze the face-to-face interview data. It 
was intended to serve the dual purposes of the project by identifying 1) building damage and 
environmental conditions before or during evacuation and 2) patterns of behavior before or during 
evacuation that may have helped or hindered occupants' evacuation. Some of the codes would apply to 
an entire interview (for example, the tower 1 , tower 2, or Building 7 code simply indicates the building 
where the respondent was located at the beginning of the event), some part of their evacuation experience 
(such as the counterflow, crowded, or elevator codes), or an observation of their surroundings (such as the 
floor damage, building damage, saw fire, or saw smoke codes). The codes are simply presented in 
alphabetical order. 

Alarm: Heard fire or other alarm (not public address system) 

Alternative Activity: Decision to suspend evacuation in favor of another activity (call, get 
drink/food/possession ...) 

Antisocial Behavior: Any directly observed behavior which presents obstacle to people getting out 

Assist: Assisted or motivated an injured / handicapped / reluctant person 

Bldg Employee: Building service employee (janitor, elevator operator, contract service employee — not 
building security) as an information source 

Boss or higher: Boss, manager, or higher as an information source 

Building 7: Occupant was in Building 7 when Tower 1 was struck 

Building Damage: Directly observed damage to building structure after beginning evacuation (see floor 
damage for damage on floor where they began their evacuation) 

Building Security: Uniformed building security personnel (not Port Authority, NYPD, or FDNY) 

Cellular Phone: Use of a cellular phone as an information source 

Concourse: Occupant mentions being in the Concourse during their evacuation. Autocoded with phrases 
such as Concourse Borders shopping mall,... 

Confusion: Unaware of where to go because of a lack of visual certainty 

Counterflow: Movement of people or responders against the flow of egress delayed evacuation 

Coworker: Coworker as an information source 

Crowded: Egress slowed because of density, but no specific identifiable cause 

Elevator - Considered: Occupant considered using elevators but did not make use of elevators for 
evacuation. Code should also be used when occupant states they knew not to use elevators. 

NIST NCSTAR 1-7, WTC Investigation 211 



Appendix B 

Elevator - Used: Used an elevator during evacuation 

Email: Used email or PDA device as an activity 

Email/PDA: Use of email or PDA as an information source 

Evacuation Decision: Marks when respondent decided they needed to leave. Note that his may occur 
more than once if the occupant decides to interrupt their evacuation on it has begun. 

Experience Aid: Prior experience aided evacuation 

Experience Obstacle: Experience caused them to delay evacuation 

FBI: Occupant specifically mentioned seeing someone identified with the Federal Bureau of Investigation 

FDNY: Fire department personnel as an information source 

Fire Affecting Egress: Observed fire in egress path 

Firefighting by Occupants: Occupant engaged in firefighting activities prior to or during evacuation 

First Awareness: First indication that respondent became aware that something was wrong on the 
morning of September 1 1 

Floor Damage: Directly observed damage to building structure on floor where they began evacuation 

prior and to beginning evacuation (see building damage for damage during evacuation and not on 
floor where they began their evacuation) 

Floor Start: Starting floor for evacuation that is different from the employer assigned floor 

Floors 1 to 9: Occupant began their evacuation in the stated range of floors 

Floors 10 to 19: Occupant began their evacuation in the stated range of floors 

Floors 100 to 110: Occupant began their evacuation in the stated range of floors 

Floors 20 to 29: Occupant began their evacuation in the stated range of floors 

Floors 30 to 39: Occupant began their evacuation in the stated range of floors 

Floors 40 to 49: Occupant began their evacuation in the stated range of floors 

Floors 50 to 59: Occupant began their evacuation in the stated range of floors 

Floors 60 to 69: Occupant began their evacuation in the stated range of floors 

Floors 70 to79: Occupant began their evacuation in the stated range of floors 

Floors 80 to 89: Occupant began their evacuation in the stated range of floors 

Floors 90 to 99: Occupant began their evacuation in the stated range of floors 

Floors below ground: Occupant began their evacuation in the stated range of floors 

Followed Crowd: Respondent reports following a crowd, or else avoiding a crowd, eg a stair looks too 
crowded. 

Fruin: An indication of Fruin level of service for crowding in the stairwells 

Get Information: Looking for more information (not milling) 

Get Out Instruction: Identifying the instruction to evacuate the building - it is of interest as to who this 
instruction came from (secretary or supervisor, for instance). 

Health: Occupant reports being tired, sick, in pain ... effecting egress 



212 NIST NCSTAR 1-7, WTC Investigation 



Qualitative Analysis Coding 



Helping: Stopped or slowed by helping other who needed assistance 

High: Floors 77 to 1 10 in the Towers and 26 to 47 in Building 7 

Immediate Evacuation: Where occupant made an immediate decision to evacuate 

Instruction: Instruction to take action counter to continuing or commencing evacuation (any source). 

Instruction Aid: Received an instruction (any source) that aided evacuation 

Interesting behavior: This code is a chance to identify any interesting behavior to highlight for the 
report. 

Internal Aids: Any thoughts or techniques which calmed occupant during evacuation (counting, 
humming, thinking of family ...) 

Know 1993: While not there, occupant was influenced by second-hand knowledge of the 1993 
WTC bombing 

Left Stairwell: Occupant left stairwell during evacuation 

Location: Indicates a specific location within a building before or during evacuation 

Location - Concourse: Indicates the occupant was on the Concourse Level 

Locked/ Jammed Door: First hand observation of a locked or jammed door effecting egress 

Low: Floors 1 to 40 in the Towers and 1 to 25 in Building 7 

Media: Use of the media (TV, radio ...) as an information source 

Medical Disability: Asthma, illness, pregnancy 

Medium: Floors 43 to 74 in the Towers 

Mill: Talked with other people inside the building as an activity 

Milling Aid: Milling as an aid to evacuation 

No alarm: Occupant specifically reports not hearing an alarm at any time during their evacuation 

Other Activity: Activities of interest to the NIST investigation not otherwise categorized 

Other Egress Aid: Aid to egress of interest to the NIST investigation not otherwise categorized 

Other Info: Other information source of interest to the NIST investigation not otherwise categorized 

Other Mobility Challenge: Mobility-impaired information of interest to the NIST investigation not 
otherwise categorized 

Other Obstacle: Obstacle effecting egress not otherwise categorized 

Overweight: Overweight or out of shape 

Pager: Occupant reported using a pager at some time before or during their evacuation 

Phone: Used a phone to receive information (not cellular phone) 

Phone Call: Respondent made use of a telephone as an activity (not cell phone). Typically to inform 

someone of status ~ family, 911, etc. For phone as an information source, use cellular phone or 
phone 

Photoluminscent Paint: Photoluminescent paint mentioned as an aid to evacuation 

Physical Assistance: Received physical assistance which aided evacuation 



NIST NCSTAR 1-7, WTC Investigation 213 



Appendix B 

Physical Disability: Wheelchair, crutches, cane, bad knees, ... 

Police: NYPD or other official gun-carrying officers as an information source 

Port Authority: PAPD or person with building responsibility as an information source 

Practice Drill: Participated in practice drills at WTC 

Present 1993: Occupant was present at the 1993 WTC bombing 

Prior Evac Experience: Participated in previous building evacuation (WTC or not), except for the 1993 
bombing. 

Prosocial Behavior: Letting people in line, crowd calming conversations, singing, counting out loud, 
emergent leadership or other behaviors which aid evacuation 

Public Address Aid: Information from public address system assisted evacuation 

Public Address System: Public address system as an information source 

Quote: A short quotation that may warrant inclusion in the final report because it creates, supports, or 
refutes a hypothesis or is particularly compelling account of a particular event 

Rest on stairs: This code identifies when the respondent actually rests in the stairs and/or observes this 
behavior. 

Retrieve: Retrieved a personal or business belonging 

Risk - High Level: High perception of danger faced by the occupant 

Risk - Low Level: Low perception of the danger faced by respondent 

Saw Airplane: Directly observed airplane hitting occupant's building (saw other impact for observation 
of plane hitting other building) 

Saw FDNY: Respondent mentions the presence of firefighters during their egress whether or not they 
helped or hindered egress or gave instructions 

Saw Fire: Directly observed flames within the building before or during evacuation (does not include 
observation of fire on exterior of another building) 

Saw Injured: Directly observed an injured or dead person, not including themselves 

Saw Other: Saw something relevant to NIST investigation not otherwise categorized 

Saw Other Impact: Directly observed impact of plane into a building other that the one the respondent 
was in (for example someone in T2 seeing plane hit Tl) 

Saw People: Observed someone doing something noteworthy, but it did not impact egress 

Saw Smoke: Directly observed smoke within the building before or during evacuation (does not include 
observation of smoke on exterior of another building) 

Saw Water: Directly observed water within the building before or during evacuation (not drinking water) 

Saw/Felt/Heard Impact: Indication of the impact of a plane hitting the building. (If you saw the impact 
of the plane hitting the other building and felt it in your building, it would get coded twice; once 
with this code and once with Saw other impact) 

Search: Searched for a person or people 

Shoes: Poor footwear choice or other's footwear cluttering pathway effects egress 

Signage: Signs posted in evacuation route aided evacuation 



214 NIST NCSTAR 1-7, WTC Investigation 



Qualitative Analysis Coding 



Slow Occupant: A slow occupant ahead of respondent in line impeded egress 

Smell/Saw Fuel: Direct observation or smell of jet fuel in the building 

Smoke Affecting Egress: Observed smoke in egress path 

Stair Unknown: Direct mention that respondent did not know which stairwell was used for evacuation 

Stairway from basement: Occupant used stairs from below-ground levels during evacuation 

Stairwell A: Occupant was in Stairwell A 

Stairwell A or C : Occupant was in Stairwell A or C (narrow stairwell) 

Stairwell B: Occupant was in Stairwell B 

Stairwell C: Occupant was in Stairwell C 

Strange: For something that just doesn't make sense and is worth further investigation 

Superflow: Prosocial behavior of letting higher priority evacuees (i.e., injured) pass faster delayed 
evacuation 

Superflow Aid: By helping someone injured or handicapped, was able to egress faster than they 
otherwise would have 

T2 Collapse: Note by respondent of collapse of Tower 2 

T2 hit: Note by respondent of Tower 2 being hit 

Time: Indicates a relative or absolute time 

Tower 1: Occupant was in Tower 1 when Tower 1 was struck 

Tower 2: Occupant was in Tower 2 when Tower 1 struck 

Training Aid: Prior training aided evacuation 

Transfer Floor: Occupant notes use of a transfer floor during egress by stairwell 

Trapped: Occupant was trapped prior to or during evacuation 

Victim: Mention of people who died in the event 

Visibility: low light condition or dust (but not smoke) effecting egress 

Visitor: Respondent specifically mentions being a visitor or being with a visitor at some point during 
their evacuation 

Walkie-Talkie: Respondent mentions the use of a walkie-talkie either by the respondent or by someone 
else 

Warden: Activities associated with responsibilities of the floor wardens 

Warden Info: Floor warden as a source of information 

Water: Water within the building before or during evacuation (not drinking water) was an obstacle to 
evacuation 



NIST NCSTAR 1-7, WTC Investigation 215 



Appendix B 



This page intentionally left blank. 



216 NIST NCSTAR 1-7, WTC Investigation 



Appendix C 

Causal Modeling 



The telephone interview results permit rigorous statistical analysis of individual components (questions). 
In order to understand what factors impacted the overall evacuation time of the average occupant in 
World Trade Center (WTC) 1 or WTC 2, two primary dependent variables were predicted: how long an 
individual delayed initiating their evacuation, where initiation is defined as entering a stairwell or elevator 
with the intention of exiting the building; and how long an individual spent traversing the stairwells. The 
sum of these two times was the total evacuation time. Multivariate regression modeling was utilized to 
rigorously establish factors which contributed to increasing the overall evacuation time. 

C.l PREDICTING EVACUATION DELAY IN THE WORLD TRADE CENTER 

This project analyzed the factors (variables) and social processes (the major paths of causal links between 
variables) that influenced people delaying the initiation of their evacuation out of WTC 1 and WTC 2 on 
September 11, 2001. Evacuation delay was defined as the number of minutes that passed from when a 
person first became aware that something was wrong until they began their evacuation. 

Method 

The purpose of the Telephone Interviews was to collect 800 computer assisted telephone interviews 
(CATI) of persons occupying either of the two WTC towers (WTC 1 and WTC 2) at the time of the 
terrorist attacks on September 11, 2001 in order to explore occupants' evacuation experiences. A total 
sample size of 800, with an allocation of 400 to each tower, was determined to simultaneously maximize 
the statistical precision within each tower. Estimates of percentages from tower specific survey data at 
400 exhibit sampling errors not greater than 2.5 percent, and 98 percent confidence intervals of 
percentages are no greater than ± 5 percent. This level of precision is more than adequate for examining 
characteristics of occupants and egress attributes. Additionally, mutlivariate modeling requires the use of 
F tests to determine significance of the regression models, i.e., testing the null hypothesis that the ratio of 
explained variance to error/residual variance is equal to or less than zero. The sample size of 400 per 
tower is more than adequate for this analysis since, for example, in a model featuring 20 independent 
variables, a sample size of 400 and a .05 level of significance (Type I error), the power of the F test to 
detect an R-squared of .06 is just over 81 percent. 

Attempts were made to equally divide the respondents among WTC 1 and WTC 2 occupants (i.e., n=400 
occupant telephone interviews from each tower). Within each of the WTC buildings, independent 
proportionate stratified samples of survivors were drawn. In other words, each occupant of a particular 
tower had an equal probability of being selected. The sample was collected and weighted in accordance 
with the analysis presented in Section 3.2.2 of this report. The questions posed to the respondents are 
included in earlier in this report. 



NIST NCSTAR 1-7, WTC Investigation 217 



Appendix C 

Pre Modeling Analysis 

Two bodies of research and theory were reviewed to inform our consideration of what explanatory 
factors/variables should be used to predict occupant evacuation from the towers. 

First, we compiled a comprehensive list of the 300 publications in the research record on human 
response/evacuation to risk events/information for natural and technological hazards and disasters, for 
example, research on floods, hurricanes, earthquake predictions, tornadoes, nuclear power plant accidents, 
hazardous chemical spills and many others. These studies included not only examinations of evacuations, 
but also events in which other protective actions were warranted. This literature was reviewed, 
summarized, and synthesized for salient theoretical constructs and the relationships between them. This 
bibliography is available in both standard and annotated formats from the Natural Hazards Research and 
Applications Information Center at the University of Colorado at Boulder. 

Second, we accessed the record of research on human evacuation during high rises fires. Experts in both 
areas of research developed a list of the most salient factors to include in this segment of our research. 

A range of key possible explanatory variables on which to gather data to use to predict evacuation delay 
in towers 1 and 2 were selected. We collected telephone interview data on these factors, and then 
statistically examined their import in predicting evacuation delay time. The list of factors follows. 

Location in Building 

Social Context 

Demographic Characteristics 

Pre-event Experience 

Roles of Responsibility 

Preparedness and Training 

Environmental Cues 

Social Cues 

Receiving Information about the Event 

Perceived Risk 

Seeking Additional Information 

Injuries 

Obtaining Help from Others 

Pre-evacuation Activities 

218 NIST NCSTAR 1 - 7, WTC Investigation 



Causal Modeling 



Preliminary statistical analyses were performed to determine which of these had and did not have salience 
in predicting evacuation delay out of WTC 1 and WTC 2. Those that had no significant impact on 
evacuation delay were excluded from further consideration. The factors that had salient impacts on 
evacuation delay were carried forward into the analysis to model tower evacuation. 

The Model That Was Estimated 

The model used to predict evacuation delay in WTC 1 and WTC 2 of the World Trade Center on 
September 11, 2001 used factors that preliminary analyses suggested as salient, closely followed general 
evacuation theory from the social and fire sciences, and is diagrammatically illustrated in Figure C-l. 
Note that e 4 - e 7 are the unexplained variance for variables 4-7, respectively. The model can be 
described as follows: (1) delay in evacuation initiation is a direct consequence of environmental cues, 
floor, obtaining information without seeking it, perceived risk, seeking additional information, and taking 
pre-evacuation actions; (2) taking pre-evacuation actions is a direct consequence of environmental cues, 
floor, obtaining information without seeking it, perceived risk, and seeking additional information; (3) 
seeking additional information is a direct consequence of environmental cues, floor, obtaining 
information that was not sought, and perceived risk; and, finally, (4) perceived risk is a direct 
consequence of environmental cues, floor, and obtaining information without seeking it. This model is 
parsimonious, consistent with the evacuation theory that stems from research on existing evacuation and 
risk communication research, and it well-represented the positive findings of our preliminary analyses of 
the many variables that could have impacted evacuation delay. 

Measurement 

Environmental cues (Xi) was measured by asking respondents about the number and type of 
environmental cues (severe signs of danger, and non-severe signs of danger) that they saw prior to 
initiating there evacuation. Answers to these questions were coded as dummy variables (0 or 1) and then 
added to form an environmental cues scale that could vary between and 2. 

Floor (X 2 ) was measured by asking respondents what floor they were on when the event started. 
Responses were coded as -7 (basement floors) to 105. Negative floor values were transposed into positive 
ones since this measure sought to determine how many floors people were from their building's floor of 
exit. Missing data was coded to the mean. 

Obtaining information without seeking it (X 3 ) was measured by asking respondents: "Now please think 
about the time period between when you first became aware that something had happened and when you 
first entered a stairwell or elevator to leave the tower. During this entire time period were you given any 
additional information about what was going on?" Answers were coded as a dummy variable where 
1 = yes and = no or missing data. 

Perceived risk (X 4 ) was measured by asking whether the occupant whether they believed that other people 
were in danger of being killed during the time when they first became aware that something had happened 
and when they first entered the stairwell or elevator to leave. Once again, answers were coded as a 
dummy variable, where 1 = yes and = no or missing data. This measure of perceived risk was used 
instead of danger to self of being killed because the latter contained insufficient variance to include in the 
analysis. 



NIST NCSTAR 1-7, WTC Investigation 219 



Appendix C 



Perceived 
Risk 



s 



/ 



Pre-Evacution 

Actions 

(x.) 



Environmental 

Cues 

(x,) 




/ 



Obtained 

Information 

(x,) 



Delay 

Initiating 

Evacuation 

(x 7 ) 




Sought 
Information 

w 



\ 



Figure C-l. Model of evacuation delay. 

Seeking additional information (X 5 ) was measured by asking the respondent whether they tried to get 
additional information about what was going on. Answers were coded a 1 = yes (for both successful and 
unsuccessful attempts to get additional information), and = no. 

Taking pre-evacuation actions (X 6 ) was measured by asking respondents whether or not they performed 
any of the following tasks: talk to another person face-to-face, gather personal items, telephone other 
people, continue working, save or transfer computer files, search for others, fight fire or smoke, move to 
another floor, or help others. These nine items were added to create a scale of taking actions (post- first 
became aware that something was wrong but pre-evacuation initiation) that varied between and 9. 

Finally, respondents were asked about evacuation delay (X 7 ) that is how many minutes passed from the 
time when they first became aware that something was wrong until they actually began their evacuation. 
Their responses were coded as an interval scale of minutes that varied between 1 to 80 minutes for 
WTC 1, and 1 to 75 minutes for WTC 2. The means, medians, and modes for WTC 1 were 5.61 minutes, 



220 



NIST NCSTAR 1-7, WTC Investigation 



Causal Modeling 



3.00 minutes, and 1 minute. These same values for WTC 2, respectively, were 6.04 minutes, 4.00 
minutes, and 5 minutes. As discussed previously, however, a significant bias for a segment of the WTC 2 
population may exist: occupants above floor 78 whose evacuation delay time exceeded 16 minutes did not 
successfully evacuate WTC 2 on September 11, 2001. NIST has estimated that approximately 75 percent 
of the occupants above the 78th floor at 8:46:30 a.m. were able to evacuate on September 11, 2001, 
therefore this bias accounts for less than 10 percent of the overall WTC 2 population. 

The Structural Equations for the Model 

The theoretical model presented in Figure 1 was represented by the following structural equations: 

X 4 = p 4 iXi + p 42 X 2 + P43X3 + e 4 (Eq.C-1) 

X 5 = p 51 X! + p 52 X 2 + P53X3 + P54X4 + e 5 (Eq. C-2) 

X 6 = p 61 X, + p 62 X 2 + p 63 X 3 + p 64 X 4 + p 65 X 5 + e 6 (Eq. C-3) 

X 7 = foX, + p 72 X 2 + p 73 X 3 + p 74 X 4 + p 75 X 5 + p 76 X 6 + e 7 (Eq. C-4) 

These equations cast perceived risk (X 4 ) as a direct linear function of environmental cues (Xi), floor (X 2 ), 
and obtained information (X 3 ). Seeking information (X 5 ) is a direct linear function of environmental cues 
(Xx), floor (X 2 ), obtained information (X 3 ), and perceived risk (X 4 ). Pre-evacuation actions (X 6 ) is a direct 
linear function of environmental cues (Xi), floor (X 2 ), obtained information (X 3 ), perceived risk (X 4 ), and 
seeking information (X 5 ). Finally, evacuation delay (X 7 ) was cast as a direct linear function of 
environmental cues (Xi), floor (X 2 ), obtained information (X 3 ), perceived risk (X 4 ), seeking information 
(X 5 ), and pre-evacuation actions (X 6 ). 

Estimation of the Model and Assessing for Regression Assumptions 

The model was estimated on the data from both WTC 1 and WTC 2. The estimated model parameters 
included path coefficients (P), explained variance for each equation (R 2 ), and other estimates. These are 
presented in Table C-l and Table C-2, respectively, for WTC 1 and WTC 2. 



The models were assessed for specification error, multicolinearity, nonlinearity, and heteroscedasticity in 
order to determine if basic regression assumptions could be reasonably made, and if the estimated model 
parameters were unbiased. 

Specification error was not determined to be a problem. The model included only major variables of 
import suggested by evacuation theory and excluded variables shown in our preliminary analyses as non- 
predictive in the data sets being analyzed. 

Multicolinearity can bias model estimates because it can increase the standard errors of estimated 
regression coefficients. The models for both WTC 1 and WTC 2 were assessed for multicolinearity in two 
ways. First multicolinearity does not impose a problem unless it is nearly perfect. The zero-order 
correlation matrices for both models (Table C-3 and Table C-4) were inspected to determine if any 



NIST NCSTAR 1-7, WTC Investigation 221 



Appendix C 





Table C-l 


Estimated 


parameter of the model 1 


for WTC 1. 




Variables" 


Path 


Equation 


Endogenous 


Exogenous 


Coefficient 


Estimate 


a 


a 


R 2 


X 4 


Xi 


P41 


.38 


.00 


.00 


.55 




X 2 


P42 


.37 


.00 








x 3 


043 


.14 


.00 




















x 5 


Xi 


Psi 


.21 


.00 


.00 


.25 




x 2 


P52 


.21 


.00 








x 3 


P53 


.01 


N/S 








x 4 


P54 


.15 


.02 




















x 6 


Xi 


P61 


.27 


.00 


.00 


.68 




x 2 


P62 


.41 


.00 








x 3 


P« 


.05 


.06 








x 4 


P64 


.08 


.04 








x 5 


P65 


.20 


.00 




















x 7 


Xi 


Pvi 


.29 


.00 


.00 


.49 




x 2 


P72 


-.17 


.00 








x 3 


P 73 


.20 


.00 








x 4 


P74 


-.02 


N/S 








x 5 


P 75 


.10 


.01 








x 6 


P76 


.47 


.00 







a. Where X] = environmental cues, X 2 = floor, X 3 = obtained information, X 4 = perceived risk, X 5 = sought information, 
X 6 = pre-evacuation actions, and X 7 = delay in evacuation initiation. 



222 



NIST NCSTAR 1-7, WTC Investigation 



Causal Modeling 





Table C-2. 


Estimated parameters of the model for WTC 2. 




Variables" 


Path 


Equation 


Endogenous 


Exogenous 


Coefficient 


Estimate 


a 


a 


R 2 


X 4 


Xi 


041 


.30 


.00 


.00 


.59 




x 2 


P42 


.49 


.00 








x 3 


P« 


.12 


.00 




















x 5 


Xi 


Psi 


.25 


.00 


.00 


.25 




x 2 


P 5 2 


.11 


N/S 








x 3 


P53 


.07 


N/S 








x 4 


P 5 4 


.17 


.02 




















x 6 


Xi 


p61 


.20 


.00 


.00 


.69 




x 2 


P62 


.36 


.00 








x 3 


P« 


.07 


.04 








x 4 


P64 


.23 


.00 








x 5 


P65 


.17 


.00 




















x 7 


Xi 


Pvi 


.13 


.01 


.00 


.56 




x 2 


P72 


-.19 


.00 








x 3 


P 73 


.23 


.00 








x 4 


P 7 4 


.05 


N/S 








x 5 


P75 


.11 


.01 








x 6 


P76 


.51 


.00 







a. Where X] = environmental cues, X 2 = floor, X 3 = obtained information, X 4 = perceived risk, X 5 = sought information, 
X 6 = pre-evacuation actions, and X 7 = delay in evacuation initiation. 



NISTNCSTAR 1-7, WTC Investigation 



223 



Appendix C 



Table C-3. Zero-order correlation matrix for WTC 1. 





Xi 


x 2 


x 3 


x 4 


x 5 


x 6 


x 7 


x, 


1 


.68 


.25 


.67 


.45 


.71 


.59 


x 2 


- 


1 


.26 


67 


.45 


.76 


.46 


x 3 


- 


- 


1 


.34 


.17 


.29 


.37 


x 4 


- 


- 


- 


1 


.43 


.64 


.47 


x 5 


- 


- 


- 


- 


1 


.55 


.44 


x 6 


- 


- 


- 


- 


- 


1 


.64 


x 7 


- 


- 


- 


- 


- 


- 


1 



a. Where X[ = environmental cues, X 2 = floor, X 3 = obtained information, X 4 = perceived risk, X 5 = sought information, 
X 6 = pre-evacuation actions, and X 7 = delay in evacuation initiation. 



Table C-4. Zero-order correlation matrix for WTC 2. a 





Xi 


x 2 


x 3 


x 4 


x 5 


x 6 


x 7 


x t 


1 


.60 


.43 


.64 


.46 


.67 


.56 


x 2 


- 


1 


.44 


.72 


.41 


.75 


.48 


x 3 


- 


- 


1 


.46 


.30 


.47 


.55 


x 4 


- 


- 


- 


1 


.44 


.72 


.55 


x 5 


- 


- 


- 


- 


1 


53 


.47 


x 6 


- 


- 


- 


- 


- 


1 


.68 


x 7 


- 


- 


- 


- 


- 


- 


1 



a. Where X] = environmental cues, X 2 = floor, X 3 = obtained information, X 4 = perceived risk, X 5 = sought information, 
X 6 = pre-evacuation actions, and X 7 = delay in evacuation initiation. 

correlations between the regressors where around .80 or higher — this is a typical cutoff value below 
which multicolinearity does not seriously bias model estimates. An inspection of the zero-order 
correlations in Table C-3 and Table C-4 led to the conclusion that coefficients were not sufficiently high 
for multicolinearity to be a problem. Second, multicolinearity was assessed by regressing each exogenous 
variable in each equation on all other exogenous variables in that equation, and the explained variances 
for these regressions were inspected to see if any approached 1.00, which would indicate a biasing level 
of multicolinearity. This assessment also led to the conclusion that multicolinearity was not a source of 



224 



NIST NCSTAR 1-7, WTC Investigation 



Causal Modeling 



bias in the estimated parameters of the model in either the data set for WTC 1 or the data for WTC 2. 
Therefore, it was concluded that the regressors in both models were orthogonal. 

The models were then assessed to determine if the assumption of linearity could be met. Exogenous 
variables in each equation were transformed to alternative nonlinear forms, for example, the natural 
logarithm of X, the square of X, the reciprocal of X, and the square-root of X. These transformed 
variables were then correlated with each of the pre-determining and endogenous variables in both models. 
None of the correlations involving the transformed exogenous variables increased substantially beyond 
the linear correlations presented in Table C-3 and Table C-4. A visual inspection of scatter plots also 
confirmed the conclusion that relationships were linear. 

The assumption of homoscedasticity was also assessed by visual inspection of regression residuals in 
scatter plots for each relationship in both models; and it was concluded that this assumption was met. 

The observed means, standard deviations, and ranges for all of the variable included in the models for 
both towers are presented in Table C-5. 

Table C-5. Observed means, standard deviations, and ranges for variable 
measurements in WTC 1 and WTC 2. 



Variable 3 


WTC1 


WTC 2 


Mean 


SD 


Range 


Mean 


SD 


Range 


x, 


.84 


.73 


2.00 


.68 


.70 


2.00 


x 2 


41.87 


25.71 


90.00 


48.99 


30.56 


104.00 


x 3 


.11 


.32 


1.00 


.21 


.41 


1.00 


x 4 


.63 


.48 


1.00 


.67 


.47 


1.00 


x 5 


.28 


.44 


1.00 


.29 


.45 


1.00 


x 6 


2.08 


1.50 


7.00 


2.37 


1.46 


8.00 


x 7 


5.61 


8.34 


79.00 


6.04 


8.06 


74.00 



a. Where X] = environmental cues, X 2 = floor, X 3 = obtained information, X 4 = perceived risk, X 5 = sought 
information, X 6 = pre-evacuation actions, and X 7 = delay in evacuation initiation. 

Judging the Success of the Models 

The estimated parameters of the models for WTC 1 and WTC 2 (see Tables 1 and 2) revealed that the 
model had a very high degree of success in explaining evacuation initiation delay, pre-evacuation actions, 
seeking information, and perceived risk in both towers. The adjusted explained variance (R 2 ) for 
perceived risk was 55 and 60 percent in WTC 1 and WTC 2, respectively. The adjusted explained 
variance for sought information was 25 percent in both towers. Respectively, the adjusted R 2 for pre- 
evacuation actions was 68 and 69 percent for WTC 1 and WTC 2, respectively. Finally, the adjusted R 2 
for delay in evacuation for WTC 1 and WTC 2 was, respectively, 49 and 56 percent. These are 
extraordinarily high levels of adjusted explained variance to observe in a study of human evacuation and 
these R 2 s, thereby, establish the strong predictive power of the models for both towers. All of the 
equations in the models for both towers were statistically significant at the .001 level or higher. 



NISTNCSTAR 1-7, WTC Investigation 



225 



Appendix C 



Direct Effects 

With a few exceptions, similar findings emerged in both towers. This lends validity to the conclusions 
that can be drawn from the analyses. The findings are presented below. First, we consider each of the four 
equations in both models for both towers (see Figure 1), and then the models are interpreted as a whole so 
that the most significant paths of influence for each tower can be distinguished. 

Predicting perceived risk. The findings that emerged regarding predicting the risk that people perceived 
were virtually identical across the two towers. The R 2 for perceived risk was 55 percent in WTC 1 and 
60 percent in WTC 2 (see Tables 1 and 2). In WTC 1, both environmental cues and floor had strong and 
similar impacts on predicting perceived risk, respectively, p 4) and p 42 were .38 and .37, while obtained 
information had a weaker but statistically significant impact, p 43 was . 14. In WTC 2, once again, both 
environmental cues and floor had strong impacts on predicting perceived risk, respectively p 4 i and p 4 2 
were .30 and .49, and, once again, obtained information had a weaker but statistically significant impact, 
P« was .12. These findings suggest that the risk that people perceived before they began their evacuation 
increased largely as a function of floor height and being exposed to environmental cues. It is most likely 
that floor had this effect due to the perceived increased time needed to evacuate as a result of being higher 
in the towers. Environmental cues likely had this effect since seeing, hearing, feeling, and so on physical 
cues that indicate danger make discounting danger — most people's natural inclination—harder to achieve. 
Obtained information likely increased perceived risk because people learned more about the seriousness 
of the event through the information they obtained. Clearly, however, information had a lesser impact on 
risk perception than did the two more salient variables of experiencing environmental cues and floor 
height. The only difference in findings between the towers was that, in WTC 2, floor height was by far 
the strongest predictor of perceived risk. 

Predicting seeking information. Once again, the findings that emerged for predicting seeking 
information were, almost, identical across both towers. Explained variance (R 2 ) for seeking information 
was 25 percent in both towers. In WTC 1, environmental cues and floor both had the strongest and 
identical impacts on seeking information, p 5[ and p 5 2 were both .21; obtained information had no impact 
on seeking information, p 53 was not statistically significant; and perceived risk had a slight impact on 
seeking information, p 54 was .15. In WTC 2, environmental cues had the strongest impact on seeking 
information, p 5[ was .25; the impacts of floor and obtained information, p 52 and p 53 , were not significant; 
and perceived risk had a slight impact on seeking information, p 54 was .17. Seeking information in times 
of rapid onset emergencies is a typical human response since people need to interpret and make sense out 
of an event before they act on it. The finding that environmental cues were the strongest predictor of 
seeking additional information is consistent with this theoretical finding about "milling" from past 
research. Obtained information had no impact on seeking information in either tower. This was likely 
because information to make sense out of the event had already been obtained. Perceived risk had a 
similar effect on seeking information — albeit lesser of an effect than environmental cues — in both towers, 
likely because it increased the urgency people had to interpret the situation. Interestingly, floor height (or 
distance from the exit) had a significant effect on seeking information in WTC 1, but not in WTC 2. 
Evacuation theory would predict that this effect would be present, it was in the tower that was struck first, 
and it was not present in the tower stuck second likely because the event began for the occupants of 
WTC 2 long before the second plane struck their tower. 



226 NIST NCSTAR 1-7, WTC Investigation 



Causal Modeling 



Predicting pre-evacuation actions. The explained variance (R 2 ) for taking pre-evacuation actions in 
WTC 1 was 68 percent, and it was 69 percent in WTC 2 (See Tables 1 and 2). In both towers, the 
strongest predictor of taking pre-evacuation actions was floor, p 6 2 in WTC 1 was .41 and it was .36 in 
WTC 2. Environmental cues was also predictive of pre-evacuation actions, p 6 i, respectively, in WTC 1 
and WTC 2 was .27 and .20, respectively. Once again, observing clues that one is at risk and being high 
in the building with a longer path to safety emerged as strong predictors, in this case of taking actions to 
ready to leave. Obtaining information had virtually no impact in either tower; p 63 was .05 in WTC 1 and 
.07 in WTC 2. Seeking information impacted pre-evacuation action, p 65 was .20 in WTC 1 and it was 
.17 in WTC 2; likely because the information obtained supported the need to evacuate and, hence, related 
to getting ready to leave. Finally, the impact of perceived risk on taking pre-evacuation actions (p 6 4) was 
.23 in WTC 2, but it was weaker in WTC 1 where is was .08. 

Predicting delay in evacuation. Explained variance (R 2 ) in evacuation delay was 49 percent in WTC 1 
and 56 percent in WTC 2. The impacts of environmental cues (p 7i ), floor (p 72 ), obtained information (p 73 ), 
perceived risk (P74), sought information (p 75 ), and pre-evacuation action (p 76 ) on delay in evacuation 
initiation, respectively, were .29, -.17, .20, -0.2, .10, and .47 for WTC 1; and for WTC 2 they were 
.13, -.19, .23, .05, .11, and .51. The greatest predictor of evacuation delay in both towers was taking pre- 
evacuation actions. Obviously, doing anything before initiating evacuation — including things to ready to 
leave — delayed departure. Setting this factor aside, some clear differences emerged between the two 
towers in terms of the relative impacts of the remaining variables in the model. Perceived risk (P 74 ) had 
no direct effect on evacuation initiation delay. This finding is consistent with general evacuation theory 
where perceived risk's impact on actual behavior is indirect through other factors. The three factors with 
the strongest direct effects on evacuation delay were the same in both towers. These were environmental 
cues, floor, and obtained information. In both towers, floor's effect was negative, that is, the more floors 
one was from the exit, the quicker people were to initiate their evacuation. Environmental cues and 
information that was received passively both increased delay in the initiation of evacuation. Finally 
seeking additional information had a minimal impact on evacuation delay. 

Paths of Greatest Influence and Conclusions 

Although each of the aforementioned findings are interesting in and of themselves, perhaps the most 
important findings that we can offer are those that emerge when all of the individual findings offered 
above are brought together and viewed at the same time in the context of the entire model. 

Bias. As discussed previously, any conclusions about evacuation initiation delay time in WTC 2 should 
consider the impact of disproportional decedent location, particularly as a source of the disproportionality 
may be highly correlated to the variable of interest, evacuation delay. In other words, those who 
exhibited long delay times in one region of the building were unable to be interviewed, thus artificially 
shortening the average delay time for one-third of the building. In the causal modeling, this would affect 
the relationship between 'floor' and 'delay initiating evacuation,' likely tending towards zero a slightly 
negative estimate (-0.19) of the beta value between the two variables. As floor was not a primary path 
which directly predicted evacuation initiation delay in WTC 2, the impact of this bias was considered 
secondary. The effect of this bias as it worked through other variables was not considered. 



NIST NCSTAR 1-7, WTC Investigation 227 



Appendix C 



WTC 1. Although there were other factors that had some lesser impacts on influencing what people did, 
the paths of causal influence that defined the main process that led to delay in the evacuation of WTC 1 
on September 11, 2001, follows. 

Environmental cues (information from the physical environment that something was terribly wrong) and 
floor (increased distance to safety) caused people to set out to find additional information, most likely 
information about what was going on and what they should do. Next, the act of seeking additional 
information, that is, "milling" about to make sense out of the situation, led people to take actions to 
prepare to evacuate. Finally, taking those actions to prepare to evacuate delayed the initiation of actually 
evacuating. 

In addition to this four step causal process, environmental cues and floor also had indirect impacts on 
evacuation delay as follows. Both factors increased the odds of seeking information and both factors 
increased the chances that people would take pre- evacuation actions prior to evacuating. Both factors also 
had direct impacts on actual evacuation delay. Environmental cues increased delay while floor decreased 
delay. 

WTC 2. Although there were other factors that had lesser influence on what people did, the paths of 
causal influence that defined the main process that led people to delay in the evacuation of WTC 2 on 
September 11, 2001 were identical to WTC 1 with one decided difference. 

Environmental cues (information from the physical environment that something was terribly wrong) and 
floor (increased distance to safety) predicted perceived risk. Environmental cues, floor, and perceived 
risk caused people to set out to find additional information. Next, the act of seeking additional 
information, that is "milling" about to make sense out of the situation, and perceived risk both led people 
to take actions to prepare to evacuate. Finally, taking those actions to prepare to evacuate delayed the 
initiation of actually evacuating. 

In addition to this five step causal process, environmental cues and floor also had indirect impacts on 
evacuation delay. Both factors increased the odds of seeking information and both factors increased the 
chances that people would take pre-evacuation actions prior to evacuating. Both factors also had direct 
impacts on actual evacuation delay. Environmental cues increased delay, while floor decreased delay. 

C.2 Predicting Normalized Stairwell Evacuation Time In WTC 1 On 

September 11, 2001 

This project analyzed the factors and social processes that influenced the normalized stairwell evacuation 
time per story of stairs for the people who evacuated out of WTC 1 on September 11, 2001. WTC 2 was 
excluded from this analysis because evacuees used stairs, elevators, and/or a combination of both for their 
evacuation. Evacuation time was defined as the average number of seconds per story of stairs that it took 
people from the time they entered a stairwell until they completed their evacuation out of the building. 
For a discussion of the data collection method, see Section 10.2.1 of this report. 



228 NIST NCSTAR 1-7, WTC Investigation 



Causal Modeling 



Pre-Modeling Analysis 

The model estimated to predict normalized stairwell evacuation time was specified as follows: 

First, we accessed the record of research on human evacuation during high rises fires. A group of 
individuals selected for the expertise on human evacuation during high rise fires was assembled. The 
individual experts developed the final list of variables, and the relationships between them, to use to 
specify the model used in this analysis. 

Second, preliminary statistical analyses were performed to determine which variables thought to be 
important predictors of normalized stairwell evacuation time by the team of experts had and did not have 
predictive salience in WTC 1. Those that had no significant impact on normalized stairwell evacuation 
time were excluded from further consideration. The factors that had salient impacts on normalized 
stairwell evacuation time were carried forward into the model used in this analysis. 

The Model that was Estimated 

The model used to predict normalized stairwell evacuation time in WTC 1 of the World Trade Center on 
September 11, 2001, is diagrammatically illustrated in Figure C-2. This model can be described as 
follows: (1) normalized stairwell evacuation time is a direct consequence of floor, evacuation decision 
delay, environmental cues, emergency responders, crowding, and evacuation interruption; (2) evacuation 
interruption is a direct consequence of floor, evacuation decision delay, environmental cues, emergency 
responders, and crowding; (3) crowding is a direct consequence of floor, evacuation decision delay, 
environmental cues, and emergency responders; (4) emergency responders is a direct consequence of 
floor, evacuation decision delay, and environmental cues; (5) environmental cues is a direct consequence 
of floor and evacuation decision delay; and, finally, (6) evacuation decision delay is a direct consequence 
of floor. This model is parsimonious, consistent with the input received from high rise fire evacuation 
experts, and it well-represented the positive findings of our preliminary analyses of the many variables 
that could have impacted normalized stairwell evacuation time. 

Measurement 

Since multiple regression analysis requires that all responses be analyzed in numerical form, answers had 
to be 'coded' or converted to numbers. By convention, = 'no', and 1 = 'yes.' Further, 2 is always a 
greater quantity of whatever the measurement is evaluating than 1. 

Floor (Xi) was measured by asking respondents which floor they were on when the event started. 
Responses were coded from negative seven (basement floors) to 105. Negative floor values were 
transposed into positive ones since this measure sought to determine how many floors people were from 
their building's floor of exit. Missing data was coded to the mean. 

Evacuation decision delay (X 2 ) was measured by asking respondents how minutes passed from the time 
when they first became aware that something was wrong until they actually began their evacuation. Their 
responses were coded as an interval scale of minutes. 



NIST NCSTAR 1-7, WTC Investigation 229 



Appendix C 



e 4 



Evacuation 
Decision 



Emergency 
Responders 



Evacuation 
Interruption 




Normal 

Stairwell 

Evacuation 

Time 



Environmental 
Cues 

Xj 



Crowding 
x s 



Figure C-2. Model of normalized stairwell evacuation time. 

Environmental cues (X 3 ) was measured by asking respondents about the number and type of 
environmental cues (severe signs of danger, and non-severe signs of danger) that they saw prior to 
initiating their evacuation. Answers to these questions were coded as dummy variables (0 or 1) and then 
added to form an environmental cues scale that could vary between and 2. 

Emergency responders (X 4 ) was measured by asking respondents whether items from a list made their 
evacuation more difficult. Affirmative answers to the "firefighters/police" choice were coded as 1, and 
"no" answers and missing data were coded as 0. 

Crowding (X 5 ) was measured by asking respondents whether items from a list made their evacuation 
more difficult. "Yes" answers to "crowded stairwells" were coded as 1, and "no" answers and missing 
data were coded as 0. 

Evacuation interruption (X 6 ) was measured by asking respondents whether they turned back at any time 
during their evacuation. "Yes" answers were coded as 1 , and "no" answers and missing data were coded 
asO. 

Finally, normalized stairwell evacuation time (X 7 ) was measured by asking respondents how much time 
passed between the moment they first began their evacuation until they exited the tower. Answers were 
coded on an interval scale; "don't know" and "refused to answer" responses were coded to the mean. 
These raw numbers were then normalized. The hydraulic model of people movement dominates the 
average evacuation time per floor, so it had to be removed from normalized stairwell evacuation time so 
as to fully explain the fluctuations (denoted by a prime) in reported evacuation time. This was 



230 



NIST NCSTAR 1-7, WTC Investigation 



Causal Modeling 



accomplished by a least squares analysis of reported evacuation time against floor for each tower, with 
statistical outliers removed. The final WTC 1 estimate was: 

AETF'j = RETj - .056; (Eq. C-5) 

where: 

AETF' = Average fluctuation in evacuation time per floor for occupant j, 
i = Floor evacuation began from for occupant j, and 

RET = Reported evacuation time for occupant j. 

The Structural Equations for the Model 

The theoretical model presented in Figure 1 was represented by the following structural equations. 

X 2 = p 21 X! + e 2 (Eq. C-6) 

X 3 = p 3 iXi + p 32 X 2 + e 3 (Eq. C-7) 

X 4 = p 4 iXj + p 42 X 2 + p 43 X 3 + e 4 (Eq. C-8) 

X 5 = p 51 X! + p 52 X 2 + P53X3 + P54X4 + e 5 (Eq. C-9) 

X 6 = p 61 Xi + p 62 X 2 + p 63 X 3 + p 64 X 4 + p 65 X 5 + e 6 (Eq. C-10) 

X 7 = p 71 X! + p 72 X 2 + p 73 X 3 + p 74 X 4 + p 75 X 5 + p 76 X 6 + e 7 (Eq. C-l 1) 

These equations cast evacuation decision delay (X 2 ) as a direct linear function of floor (Xi). 
Environmental cues (X 3 ) was set as a direct linear function of floor (Xi) and evacuation decision delay 
(X 2 ). Emergency responders was seen as a direct linear function of floor (Xj), evacuation decision delay 
(X 2 ), and environmental cues (X 3 ). Crowding was cast as a direct linear function of floor (Xi), evacuation 
decision delay (X 2 ), environmental cues (X 3 ), and emergency responders (X 4 ). Evacuation interruption 
(X 6 ) was cast as a direct linear function of floor (Xi), evacuation decision delay (X 2 ), environmental cues 
(X 3 ), emergency responders (X 4 ), and crowding (X 6 ). Finally, normalized stairwell evacuation time was 
cast as a direct linear function of floor (Xi), evacuation decision delay (X 2 ), environmental cues (X 3 ), 
emergency responders (X 4 ), crowding (X 5 ), and evacuation interruption (X 6 ). 

Estimation of the Model and Assessing for Regression Assumptions 

The model was estimated on the data from WTC 1. The estimated model parameters included path 
coefficients (betas), explained variance for each equation, and other estimates. These are presented in 
Table C-6. The zero-order correlations between each of the variables are presented in Table C-7. 



NIST NCSTAR 1-7, WTC Investigation 23 1 



Appendix C 





Table C-6. 


Estimated parameters of the model for WTC 1. 




Variables" 


Path 


Equation 


Endogenous 


Exogenous 


C oeff icient 


Estimate 


a 


a 


R 2 


X 2 


Xi 


B 2[ 


.58 


.00 


.00 


.34 
















x 3 


X, 


B 3] 


.78 


.00 


.00 


.79 




x 2 


B 32 


.17 


.00 




















x 4 


Xi 


041 


.23 


.00 


.00 


.57 




x 2 


042 


-.01 


N/S 








x 3 


P« 


.56 


.00 




















x 5 


Xi 


Psi 


.18 


.00 


.00 


.72 




x 2 


P 5 2 


.03 


N/S 








x 3 


P53 


.37 


.00 








x 4 


P54 


.35 


.02 




















x 6 


x, 


Pel 


.12 


N/S 


.00 


.11 




x 2 


P62 


.04 


N/S 








x 3 


P 63 


.15 


N/S 








x 4 


P64 


.01 


N/S 








x 5 


P65 


.07 


N/S 




















x 7 


x, 


P71 


.03 


N/S 


.00 


.44 




x 2 


P72 


-.05 


N/S 








x 3 


P73 


.46 


.00 








x 4 


P74 


.09 


N/S 








x 5 


P 75 


.08 


N/S 








x 6 


P76 


.18 


.00 







a. Where X] = floor, X 2 = evacuation delay, X 3 = environmental cues, X 4 = emergency responders, X 5 = crowding, X 6 ; 
evacuation interruption, and X 7 ^normalized stairwell evacuation time. 



232 



NIST NCSTAR 1-7, WTC Investigation 



Causal Modeling 



Table C-7. Zero-order correlation matrix for WTC 1. 





x, 


X 2 


x 3 


x 4 


x 5 


x 6 


x 7 


X, 


1 


.58 


.88 


.71 


.77 


.84 


58 


x 2 


- 


1 


.63 


.47 


.53 


.25 


.34 


x 3 


- 


- 


1 


.75 


.81 


.35 


.58 


x 4 


- 


- 


- 


1 


.77 


.28 


.54 


x 5 


- 


- 


- 


- 


1 


.31 


.57 


x 6 


- 


- 


- 


- 


- 


1 


.39 


x 7 


- 


- 


- 


- 


- 


- 


1 



a. Where X] = floor, X 2 = evacuation delay, X 3 = environmental cues, X 4 = emergency responders, X 5 = crowding, X 6 = 
evacuation interruption, and X 7 ^normalized stairwell evacuation time. 

The model was assessed for specification error, nonlinearity, and heteroscadasticity in order to determine 
if basic regression assumptions could be reasonably made, and if the estimated model parameters were 
unbiased. The multicolinearity assumption did not have to be assessed since the model contained only one 
exogenous variable. 

Specification error was determined not to be a problem. The model included only variables of import 
suggested by individual experts, and excluded variables shown in our preliminary analyses as non- 
predictive in the data sets being analyzed. 

The model was then assessed to determine if the assumption of linearity could be met. Exogenous 
variables in each equation were transformed to alternative nonlinear forms, for example, the natural 
logarithm of X, the square of X, the reciprocal of X, and the square-root of X. These transformed 
variables were then correlated with each of the pre-determining and endogenous variables in the model. 
None of the correlations involving the transformed exogenous variables increased substantially beyond 
the linear correlations presented in Table C-7. A visual inspection of scatter plots also confirmed the 
conclusion that relationships were linear. 

The assumption of homoscedasticity was assessed by visual inspection of regression residuals in scatter 
plots for each relationship in both models, and it was concluded that this assumption was met. 

The observed means for each of the variables contained in the model were: 47.2 for floor, 5.36 for 
evacuation delay, 1.73 for environmental cues, 0.60 for emergency responders, 0.73 for crowding, 0.12 
for evacuation interruption, and 13.01 for normalized stairwell evacuation time. 



NISTNCSTAR 1-7, WTC Investigation 



233 



Appendix C 



Judging the Success of the Models 

The estimated parameters of the model for WTC 1 (Table C-6) revealed that the model had a very high 
degree of success in explaining normalized stairwell evacuation time. The adjusted explained variance 
(R 2 ) for normalized stairwell evacuation time was 44 percent, 1 1 percent for evacuation interruption, 72 
percent for crowding, 57 percent for emergency responders, 79 percent for environmental cues, and 34 
percent for evacuation initiation delay. With the exception of evacuation interruption, these are 
extraordinarily high levels of adjusted explained variance to observe in a study of human evacuation. 
These R 2 's, thereby, establish the strong predictive power of the model. All of the equations in the model 
were statistically significant at the .001 level or better. 

Direct Effects in the Model 

We first consider each of the six equations in the model (see Figure 1), and then the model is interpreted 
as a whole so that the most significant paths of influence can be distinguished. 

Predicting evacuation initiation delay . The findings that emerged regarding predicting delay in the 
initiation of evacuation from floor were that the R 2 was 34 percent with a value of .58 for p 2 i significant at 
the .001 level. This relationship was elaborated earlier in this appendix. 

Predicting environmental cues . Explained variance (R 2 ) for observing environmental cues was 
79 percent, and the equation was statistically significant at the .001 level. Floor had a very strong direct 
impact on observing environmental cues; p 3i was .78 and it was significant at the .001 level. The effect of 
delay in the initiation of evacuation (p 32 ) was .17, and it was also statistically significant at the .001 level. 
It appears that the longer a person took to begin their evacuation, the more the physical impacts of the 
event grew and the more likely people were to experience them. 

Predicting emergency responders . The explained variance (R 2 ) for the third equation in the model 
predicting encountering emergency responders was 57 percent, and the equation was statistically 
significant at the .001 level. The relative effects of floor, delay in beginning evacuation, and 
environmental cues, respectively, were as follows: p 4i was .23 statistically significant at the .001 level, P42 
was -.01 and was not statically significant, and p 43 was .56 statistically significant at the .001 level. Floor 
height and experiencing environmental cues both predicted encountering emergency responders. This 
makes sense when one considers that emergency responders would be most likely to go to areas 
experiencing the impacts that would also yield environmental cues, and the higher one was in the tower, 
the more stairwells one had to traverse and the greater the odds of encountering emergency responders. 

Predicting crowding . The explained variance (R 2 ) for the fourth equation in the model that predicted 
perceived crowding on the evacuation stairwells was 72 percent. The four variables in the equation had 
the following effects on perceived crowding: floor (p 5 i) was .18, significant at the .001 level; evacuation 
delay (p 52 ) was .03, and it was not statistically significant; environmental cues (p 53 ) was .37, significant at 
the .001 level; and encountering emergency responders (p 54 ) was .35, significant at the .001 level. 
Perceived crowding largely increased as a result of environmental cues and encountering emergency 
responders. 

Predicting evacuation interruption . The explained variance (R 2 ) for interrupting evacuation was only 
1 1 percent. Even though this equation was statically significant at the .001 level, none of the 5 predictor 

234 NIST NCSTAR 1 - 7, WTC Investigation 



Causal Modeling 



variables in the equation had a statistically significant impact on evacuation interruption. The beta 
weights for these variables (floor, evacuation delay, environmental cues, emergency responders, and 
crowding) were, respectively (p 6 i through p 65 ) .12, .04, .15, .01, and .07. At best, we can interpret these 
findings to mean that there was a slight but not statistically significant tendency for people to interrupt 
their evacuation if they had more rather than fewer floors to traverse to safety, and if they encountered 
environmental cues (perhaps obstacles) in the process of evacuation. 

Predicting normalized stairwell evacuation time . Explained variance (R 2 ) in predicting normalized 
stairwell evacuation time was 44 percent, and the equation was statistically significant at the .001 level. 
The impacts of floor (p 7 i), evacuation delay (P72), environmental cues (P73), emergency responders (P74), 
crowding (p 75 ), and evacuation interruption (p 76 ), on normalized stairwell evacuation time, respectively, 
were .03, -.05, .46, .09, .08, and .18. Of these, only two factors in the equation were statistically 
significant both at the .001 level. These were environmental cues (p 73 was .46) and evacuation interruption 
(P 76 was .18). Clearly, the single factor that had the biggest impact on increasing the amount of time 
people spent, on average, per stairwell was environmental cues. The only other factor that had a 
significant impact was interrupting evacuation, obviously, because stopping egress would increase the 
amount of time needed to complete evacuation. 

Conclusions 

Although each of the aforementioned findings are interesting in and of their own right, the most important 
findings that we can offer are those that emerged when all of the above findings are brought together and 
viewed at the same time in the context of the model as a whole (see Figure 1). What doing so revealed 
was that, based on this analysis, the main process that led to increased normalized stairwell evacuation 
time in the evacuation of World Trade Center WTC 1 on September 11, 2001, was straightforward and 
clear. It can be described as follows. 

Floor (increased distance to safety) substantially increased the odds that people would encounter 
environmental cues. Floor also increased delay in starting evacuation, which, in turn, also increased the 
chances that people would encounter environmental cues. But it was encountering environmental cues 
(which likely blocked egress) that had a large and direct effect on increasing the amount of time that 
people spent, on average, to traverse their evacuation stairwell. In addition to this multi-step process with 
environmental cues as the key predicting variable, interrupting the process of evacuation for any reason 
also increased the amount of time, on average, that people used to descend their evacuation stairwell. 



NISTNCSTAR 1-7, WTC Investigation 235 



Appendix C 



This page intentionally left blank. 



236 NIST NCSTAR 1-7, WTC Investigation 



Appendix D 

Egress Modeling 



The purpose of modeling evacuation from the World Trade Center (WTC) towers was to obtain 
evacuation times for a variety of scenarios which provide context to understanding the actual evacuation 
on September 11, 2001. The following six scenarios were selected and modeled: 

Phased evacuation 

• Scenario 1 : Occupants in an emergency situation have to travel 3 floors below the fire floor, floor 
48 

Total evacuation 

Scenario 2: Full capacity tower without visitors or damage 

Scenario 3: Full capacity tower including visitors without damage 

Scenario 4: Sept. 1 1th capacity tower without damage 

Scenario 5: Full capacity tower with plane damage blocking floors 91 and above in WTC 1 

Scenario 6: Full capacity with plane damage blocking floors 78 and above in WTC 2 after 
16 minutes, including estimates of elevator usage 

D.l METHOD FOR THE SIMULATING PHASED EVACUATION (SCENARIO 1) 

The purpose of simulating a phased evacuation was to understand how the evacuation strategy would 
have worked during an emergency at the WTC towers. For a phased evacuation, the occupants of the 
emergency floor, the occupants on the floor above, and the occupants on the floor below were to evacuate 
to three floors below the emergency floor. On floors 47 through 49, 200 occupants were placed on each 
floor. These 600 occupants traveled to the 45th floor where they would be considered "safe." Floor 48, 
the emergency floor, was a transfer floor; therefore, occupants from floor 49 were faced with moving 
through the horizontal corridor on the 48th floor. Figure D-l shows a three-dimensional view of the 
45th - 50th floors, as entered into the buildingEXODUS model. The occupants were placed on 
floors 47 - 49 and the three stairwells connected all 6 floors together. 



NIST NCSTAR 1-7, WTC Investigation 237 



Appendix D 




Figure D-l. 3-D view of a phased evacuation showing floors 45 - 50 (Exodus). 

Figure D-2 shows occupants traveling through one of the transfer corridors on floor 48. Occupants are 
traveling on the stairs that connects floor 49 and 48, walking through the transfer corridor and eventually 
onto the stair connecting floor 48 to 47. 




Figure D-2. Transfer corridor on floor 48 (Exodus). 



D.2 



METHOD FOR SIMULATING TOTAL EVACUATION (SCENARIOS 2-6) 



Total building evacuation represented an emergency in which all occupants of the building evacuated 
simultaneously. Scenarios 2-6 were total evacuation simulations, which varied the total occupant 
population inside a tower and the damage to the building at the time of evacuation. 



D.2.1 



Population Estimates 



Scenarios 2-4 required population estimates. Scenario 2 required an estimate of the number of 
occupants in the building, without the inclusion of visitors. Furniture and egress floor plans were studied 
from a sampling of floors: in WTC 1 - floors 94-100 and in WTC 2 - floor 77. On each floor, the 



238 



NIST NCSTAR 1-7, WTC Investigation 



Egress Modeling 



number of offices, cubicles, and other workstations was counted. From these numbers, the usable floor 
area (square footage) of the floors was divided by the corresponding number of persons to obtain an 
occupant density (area per person). Each floor in the towers received an average density (14 m 2 /person), 
and the number of occupants per floor changed according to the total usable area of each floor. This 
resulted in a total of 19,800 persons to be simulated in Scenario 2. 

For scenario 3, the full-capacity population with visitors, the number of occupants inside the towers was 
determined by the same method as Scenario 2, adjusted for visitor spaces. Visitor spaces included 
conference rooms, waiting areas, lunch tables, and library chairs. The same floor plans were used from 
Scenario 2 population development to count these additional spaces. Then, as in the previous calculation, 
the usable floor area of each floor in the towers was used to calculate the number of occupants, including 
visitors in each tower. With these calculations, the occupant density was 1 1 m 2 /person when including 
visitors. This resulted in an estimate of 25,500 people. The findings of each step of the estimation 
method are summarized in Table D-l. Note, however, that in 2005, PANYNJ estimated that the 
maximum population of WTC 1 or WTC 2 would not likely have exceeded 20, 000. 68 

Table D-l. Floor and occupant modeling parameters. 









No. of chairs 


No. of chairs (with 


Zone 


Represented Floors 


Approx. area (m 2 ) 


(w/o visitors) 3 


visitors)* 


1 


9-26 

27-34 
35-40 


2,723 
2,826 
2,922 


200 


260 


2 


43-47 


2,679 


190 


240 


3 


49-54 
55-56 

57-74 


2,974 
2,935 
3,007 


210 


270 


4 


77-81 


2,860 


200 


260 


5 


83-95 
103 - 104 
96-102 
105 - 106 


3,188 
3,309 

3,245 
3,245 


230 


300 



a. Averaged over variations within each zone. 

The population for Scenarios 4, 5, and 6 was based upon the National Institute of Standards and 
Technology (NIST) estimate of the number of occupants present at WTC 1 and WTC 2 on 
September 11, 2001, or approximately 8,800. The purpose of Scenario 4 was to obtain a lower bound 
evacuation time for a September 11, 2001 population in the absence of constraints to evacuation or 
elevator usage. The 8,800 estimate was used to 'calibrate' (refine the model output to match the gross 
evacuation characteristics from the actual evacuation of WTC 1 and WTC 2) the model for Scenarios 5 
and 6, where additional occupants were added in combination with building damage (Galea 2004). 

For the total building evacuation scenarios, occupants were placed on all floors of the building, excluding 
mechanical spaces. Floors 1 through 107 were modeled (floors 108-110 were not normally occupied). 
During each simulation, the occupants left their floors and traveled into one of the three stairwells, in 



1 Bhol, Saroj. PANYNJ (September 21, 2005). Email from S. Bhol to S. Sunder in response to NIST question. 



NISTNCSTAR 1-7, WTC Investigation 



239 



Appendix D 



which they remained throughout the entire evacuation. The stairwells included all transfer corridors and 
emptied into the Mezzanine or the lobby area, depending upon the stairwell. Occupants reached safety as 
soon as they exited the tower (either to the outside or into the Concourse mall area). 

Finally, the value of 8,800 occupants for the number of occupants in a WTC tower on 

September 11, 2001 was based upon the statistical projection derived from the telephone survey data 

described previously in Table 4-1. 

D.3 WTC Tower Geometry Development 

Each Tower, without including the basement floors, contained a total of 1 10 floors. In order to ease the 
difficulty of floor plan input in each evacuation model, 1 1 representative floor plans were created and 
replicated throughout the building. The core layout and stair position did not change significantly in 
between the transfer floors. Therefore, knowing the floor location of core and stair walls, the tower was 
divided into zones, the boundaries of which were transfer floors. Either tower (since they were identical 
in core layout and stair position) was divided into 5 zones, in addition to the four significant transfer 
floors (floors 42, 48, 76, and 82). The 1 1 representative floors were the Concourse, Mezzanine, a 
representative floor from each of the 5 zones, and the 4 unique transfer floors. Using representative floor 
plans greatly simplified floor plan input into each evacuation model used, with little loss of accuracy. 

Even though floor duplication was used, each floor had to be created individually to achieve an accurate 
tower description within the evacuation model. At and around the mechanical spaces, the floor to ceiling 
measurements increased, meaning that throughout the building, the number of stair risers between floors 
ranged from 18 to 26 risers throughout the tower. Therefore, each floor was developed individually to 
ensure it contained the correct number of stairs to connect to the floor above. Also, even though 
occupants were not placed on these floors, mechanical floors were added to each model to include the sets 
of stairs that led throughout the mechanical spaces. 

Three additional stairwell movements were not modeled, as the stairwell translation was less than 5 m: 
floors 26-41 (Stair A) and 66-68 (Stairs A and C). Note that the stair movement on floors 26-41 was 
input into the buildingEXODUS evacuation model only, as stair positioning was more important in this 
model, as discussed below. 

Table D-2 shows the major floors plans contained in each zone. In Zones 1, 3, and 5, there were minor 
changes to the core walls from floor to floor, shown in the additional floor plans. The table outlines 
which floors these changes take place. However, due to the size of the building and the relative 
insignificance of a core wall change to the evacuation (since most of an occupant's travel time was spent 
in the stairwells), the floor plan of the majority of floors within a zone was used as the representative floor 
in each evacuation model. Table D-2 shows that floor plan Zone 1-1 was used as a representative of 
Zone 1, Zone 3-3 as a representative of Zone 3, and Zone 5-1 as representative of Zone 5. 



240 NIST NCSTAR 1-7, WTC Investigation 



Egress Modeling 



Table D-2. WTC floor plan with core walls and stairwell locations. 

Zone 1: Floors 3-41 



no 

o II □ 



Zone 1-1: Floors 9-26 
There is a slight change in 
Stair A at floor 25. 
**Used for all of Zone 1 




Zone 1-2: Floors 27-34 
Variations are found in the upper 
right quadrant throughout this 
zone. 




Zone 1-3: Floors 35-41 
Variations are found in the upper 
right quadrant throughout this 
zone. 



Transfer Floor 42 



b 


^^ 



Stairs A and C transfer 




View of Stair A transfer 



Zone 2: Floors 43-47 



Escalator positions: There is an escalator on the right side of the upper 
right quadrant between floors 44-45; Escalator on left side of upper left 
quadrant between floors 43-45 - these will not be used for evacuation 
purposes. 



b 


n 



Transfer Floor 48 



h 


nr In 



Stairs A and C transfer 




View of Stair C Transfer 



NISTNCSTAR 1-7, WTC Investigation 



241 



Appendix D 



Zone 3: Floors 49-75 




Zone 3-1: Floors 49 -54 




Zone 3-3: Floors 55-56 




Zone 3-3: Floors 57-75 

1) Variations are found in the 
upper left and right quadrants 
throughout this zone. 

2) There is a slight change in 
Stairs A, C between floors 66-68. 
*Used for all of Zone 3 



Transfer Floor 76 



C 






Stairs A, B, and C transfer 




View of Stair B Transfer 



Zone 4: Floors 77-81 












Q 


"1 





Transfer Floor 82 





^L 


i-j," 




tr- 



Stairs A and C transfer 




View of Stair C Transfer 



242 



NIST NCSTAR 1-7, WTC Investigation 



Egress Modeling 



Zone 5: Floors 83-107 



uu 


c 





Zone 5-1: Floors 83-95 

Variations are found in the lower 

left quadrant throughout this 

zone. 

*Used for all of Zone 5 




Zone 5-2: Floors 103-104, 107 
Variations are found in the lower 
left quadrant at floor 107. 





— ^ 


u 

rV 



Zone 5-3: Floors 96-102, 105- 

106 

Variations are found in the lower 

left quadrant throughout this 

zone. 



D.3.1 Single-Tenant and Multi-Tenant Floor Plans 

The floor plans shown in Table D-2 (used in the evacuation modeling) were single-tenant (open) floor 
plans, which contained only the outer walls, the core walls, and the stairwells. This section evaluates the 
quality of the assumption that the subdivision of a floor space into multiple tenancies did not significantly 
change the model results when compared to modeling the space as a single open floor plan. The 
assumption was based upon the premise that in a building as large as WTC 1 or WTC 2, most of the 
average occupant's evacuation time would be spent in the stairwell. Specific multi-tenant floor plans 
were chosen to study, developed based upon the demising wall layouts found in the WTC Space Book 
Plan (PANYNJ 2001c). The Simulex model was used for this analysis. 

An open, single-tenant floor plan from each zone (1-5) was simulated by placing the appropriate number 
of occupants on each floor, running the model, and recording the numbers of occupants using each stair. 
The percentage of occupants who used each stairwell was calculated for each zone. Then, by factoring in 
the number of floors in each zone (that the sample floor represents), the weighted percentages of occupant 
use of each stair was calculated for the entire building. These percentages are shown in on the left hand 
side, labeled as the single-tenant floor plan percentages. 

Using the Space Book Plans, at least three representative multi-tenant floor plans were selected from each 
zone. Each multi-tenant floor plan contained at least 3 corridors, and/or at least 1 different companies. 
In each of the five zones, the movement of occupants within the multi-tenant floor plans was simulated, 
and the percentages of stair use were averaged over each zone. The number of multi-tenant floors and 
single-tenant plans per zone were calculated and assigned their appropriate average percentages. As 
shown in Table D-3, the weighted percentages of occupant use of each stairwell were calculated for the 
entire building. 



NISTNCSTAR 1-7, WTC Investigation 



243 



Appendix D 



Table D-3. Stair use percentages for single-tenant versus multi-tenant floor plans. 



Single-tenant Floor Plans only 


Single- and Multi-tenant Floor Plans 


Stairwell C 


Stairwell B 


Stairwell A 


Stairwell C 


Stairwell B 


Stairwell A 


32.96 % 


31.71 % 


35.33 % 


33.50% 


32.40 % 


34.10% 


Number of people in each stair with 19,800 population 


6526 


6279 


6995 


6633 


6415 


6752 


Difference (Gross) 


-107 


-136 


243 


Difference (Percent) 


-0.5 % 


-0.68 % 


1.2% 



In summary Table D-3 shows that there was approximately 1 percent difference between assuming all 
floors to be single tenant (open) floor and modeling both single- and multi-tenant floors with respect to 
the choice of stairwells. Therefore, representative single-tenant (open) floor plans were used in the 
subsequent evacuation modeling. 



D.4 



EVACUATION MODELS 



Three evacuation models were used. The Simulex model (IES 2001) was used to perform the single- 
tenant and multi-tenant floor plan analysis and the phased evacuation scenario (Scenario 1). The 
buildingEXODUS (Gwynne et al. 1998) model was used for all scenarios from the WTC towers 
(Scenarios 1-6). Finally, the EXIT89 model (Fahy 1999) was used for the phased evacuation and the total 
evacuation scenarios (without damage) from the WTC towers (Scenarios 1-4). This section provides brief 
descriptions of each of the three models used in this project as well as the reason for picking these three 
among all other available evacuation models. 

Simulex 

The Simulex model was developed by Thompson, from IES, Inc. in Scotland, and is a widely used model 
in the field of fire safety. The model has been validated and focuses on the movement of occupants 
throughout a structure (IES 2001), (IES 2000). The model allows the user to specify a distribution of 
occupant types, associated with a certain body size and unimpeded movement speed, throughout the 
building. The model includes important occupant movements, such as overtaking, side-stepping, back- 
stepping, and others. Simulex predicts that occupant movement throughout building spaces will slow (in 
speed) as the inter-person distance between occupants and / or obstacles decreases. The model has fixed 
limitations on the number of floors and exits that the model can handle, however the developer is willing 
to expand these to meet the needs of the users. Also, Simulex provides a 2-D visualization of the 
evacuation throughout the building and allows the user to input CAD drawings of the building for ease in 
building geometry development. 

The input required for this model included the following: 

• Floor plans in the form of CAD drawings 

• Connections of floor levels by stairways or ramps (involving user input on stair width and length) 

• Distance maps that can be created to block certain exits or paths from groups of occupants 



244 



NIST NCSTAR 1-7, WTC Investigation 



Egress Modeling 



• Occupant movement characteristics for each individual or a group of individuals with a 
corresponding body size, initial horizontal speed, and percentage decrease of speed on stairs (the 
user can use default values provided by Simulex or create his / her own) 

• Occupant delay times to be assigned from a random, triangular, or normal distribution 

building EXODUS 

buildingEXODUS (Gwynne et al 1999; 1998; 2001) was developed by Galea at the University of 
Greenwich, UK. buildingEXODUS is a well-documented and well-validated model, claimed to be 
limited only by the storage capacity of the computer used to run the program. The model contains six 
submodels that work together to form the evacuation simulation: occupant, movement, behavior, hazard, 
geometry, and toxicity submodels. Any time occupants wish to occupy the same node, or a node that is 
already occupied, a conflict resolution time is added to their total evacuation time. The method of conflict 
resolution is how the model simulates congestion throughout the building. Individual abilities of the 
occupants and conflict resolution are probabilistic. The behavioral model makes route choices depending 
upon the environmental situation of the building and then passes this information onto the movement 
model. Familiarity and awareness variables can be assigned to exits, and drive and patience variables are 
assigned to the occupant. These variables are random and difficult for the user to specify for any 
evacuation. buildingEXODUS provides the user with 2-D and 3-D visualization capabilities of the 
building evacuation and allows for the input of CAD drawings of the building for ease in building 
geometry development. 

The input required for this model included the following: 



• 



• 



• 



Floor plans in the form of CAD drawings (other methods are available to provide building 
information) 

Connections of floor levels via stairways, with an option for including effective width inside the 
stairwell 

Distribution of occupant types throughout the building (user can specify information about each 
individual or group such as speed, ability, drive, patience, etc.) 

Option of including special nodes at certain points in the building to regulate occupant paths 
through building and flow through the exits/doors 

• Option of including boundary nodes used to predict slower movement near walls, obstacles 

• A series of simulation options such as stair packing, occupant randomization throughout the 
building, response time distribution, stair edge preference, and extreme behavior 

EXIT89 

EXIT89 (Fahy 1999), (Fahy 2001) is an evacuation model developed by Fahy at the National Fire 
Protection Association (NFPA). The model has been tested on tall buildings in the past and has the 
capability of modeling a large number of occupants in a building. The model relies on the density versus 

NIST NCSTAR 1-7, WTC Investigation 245 



Appendix D 



speed data from Predtechenskii and Milinskii (Predtechenskii and Milinskii 1969) for different building 
components, such as horizontal components, doorways, up stairs, and down stairs. It also uses 
conditional movement, depending upon the presence and density of smoke in the evacuation path. All 
occupants in the building are made up of the same body size and density versus speed distribution, both 
specified by the user. EXIT89 uses a series of nodes and arcs instead of CAD drawings to represent any 
type of structure. In this type of model, the floor plan is entered as a series of nodes (rooms, corridor, 
stair sections, etc.) and arcs (distance between nodes). When using a node/arc model for a single-tenant 
(open) floor plan such as those used for the WTC Towers, the user determines how to segment and link 
the building space, and then should check to make sure that occupants are traveling in realistic patterns to 
the exits. While EXIT89 does not provide visualization capability, the model does provide a particularly 
detailed output that notes where each occupant is at each time period throughout the evacuation. EXIT89 
contains a variety of input choices for the user pertaining to each evacuation simulation: 

• Shortest travel route or user-defined route for occupants 

• The use of CFAST smoke data, user-defined blockages, or none 

• The choice of a body size for occupants which applies to entire population of the building - Large 
(0.1458 m 2 ), Medium (0.1 13 m 2 ), or Small (0.0906 m 2 ) 

• The choice of speed for the entire population of the building - Emergency (horizontal unimpeded 
speed = 1.36 m/s) or Normal (horizontal unimpeded speed = 0.91 m/s) 

• A random delay time (uniform distribution) 

• The modeling of disabled occupants, including the percentage of decrease in travel speed for 
these occupants compared with the rest of the population 

The three models described above were used in different capacities for this project. Simulex was used in 
a limited capacity to provide analyses of occupant movement on individual floors and to simulate the 
phased evacuation. EXIT89 and buildingEXODUS were used to model scenarios of an entire 
WTC Tower with a large number of occupants within the building. 

The three models were chosen for this project for different reasons. The Simulex model was chosen to 
perform certain aspects of the project due to its heavy focus on understanding and accurately predicting 
occupant movement, and was used for some of the smaller aspects of the project. The following reasons 
justified buildingEXODUS and EXIT89 as reasonable choices to model WTC 1 and WTC 2: 

• Both had the capacity to simulate evacuation from a 1 10-story building with 25,500 people. 

• Both provided the output that was needed to answer evacuation questions, including visualization 
capabilities from buildingEXODUS 

• Both could handle the complexity of the building, such as the transfer floors and differences in 
floor to ceiling heights in different part of the Tower 

• Both provided explanation for the underlying data used for the movement of occupants 
throughout the Tower 

246 NIST NCSTAR 1 - 7, WTC Investigation 



Egress Modeling 



Both provided sufficient knowledge and direction to NIST, through personal contact and/or user 
manuals, on how to use the model 



• 



• Both developers had published validation studies on the model 

• Both models had been used by NIST or outside of NIST. This way NIST could identify their 
inefficiencies ahead of time and compensate for these during this project. 

D.4.1 Phased Evacuation (Scenario 1) Inputs and Results 

All three models were used to perform the phased evacuation analysis, Scenario 1. Each model required 
building geometry, population, and behavioral input conditions. The phased evacuation scenario involved 
the simulation of floors 45-49, and the emergency floor was floor 48. All three stairs, containing 19 risers 
per floor from floors 45-49, were added to the models with the appropriate diagonal length, stair width, 
and door width into and out of the stairwell. The number of occupants used in the phased evacuation was 
200 per floor on floors 47-49, which totaled to 600 occupants. The number of six hundred occupants was 
used as a rounded estimate of the number of occupants seen on three floors within Zone 2 of the building. 
The movement response delay was set to zero for one set of simulations, and ranged between zero and ten 
minutes for the other set of simulations. 

The difference between models for the phased evacuation scenario was the population characteristics of 
the occupants. The population type used in the Simulex model was the "office staff type which 
distributed 60 percent males and 40 percent females throughout the building. The population distribution 
input into buildingEXODUS for the phased evacuation scenario was taken from WTC 1 telephone survey 
demographic data: 5 percent males age 17-29, 38 percent Males age 30-50, 21 percent Males age 51-80, 
3 percent Females age 17-29, 22 percent Females age 30-50, and 11 percent Females age 51-80. All 600 
occupants simulated with the EXIT89 model were assigned the medium body size and emergency speed 
parameters. 

The total evacuation time for the phased evacuation, Scenario 1, was similar across all three models. 
Without a time delay, Simulex predicted that occupants would travel to floor 45 in approximately 240 
seconds, buildingEXODUS predicted 243 seconds, and EXIT89 (emergency conditions) predicted 210 
seconds. With a to 10 minute time delay, Simulex predicted that occupants would travel to floor 45 in 
approximately 690 seconds, buildingEXODUS predicted 660 seconds, and EXIT89 (emergency 
conditions) predicted 690 seconds. These results are shown in Table D-4. From the three evacuation 
models, the average minimum time to complete a phased evacuation in WTC 1 or WTC 2 building was 
approximately 230 seconds, or around four minutes, assuming no evacuation delay and approximately 
680 seconds, or around eleven minutes, assuming a to 10 minute time delay. 



NIST NCSTAR 1-7, WTC Investigation 247 



Appendix D 



Table D-4. Phased evacuation model input conditions. 



Evacuation Model 


Occupant Type/ Characteristics 


Evacuation 
Time, No 
Delay (s) 


Evacuation 

Time, 

- 600 s 

Delay (s) 


Simulex 


All office staff; 60 % men, 40% women 


240 


690 


EXODUS 


WTC 1 distribution: 5 % Males age 17-29, 38 % Males age 
30-50, 21 % Males age 51-80, 3 % Females age 17-29, 22 % 
Females age 30-50, and 1 1 % Females age 51-80 


243 


660 


EXIT89 


All Medium body size and emergency speed 


210 


690 



D.4.2 Scenarios 2 and 3: Inputs and Results 

Two models, buildingEXODUS and EXIT89, were used to perform the total building evacuation analysis 
of a full-capacity tower with and without visitors, Scenarios 2 and 3. Each model required building 
geometry, population, and behavioral input conditions. 

buildingEXODUS 



The buildingEXODUS model was used to simulate Scenarios 2 and 3, a total building evacuation of a 

fully-occupied tower both with and without the inclusion of visitors. Without visitors, the population 

totaled 19,800 people and with visitors, the population totaled 25,500 people. As discussed earlier, the 

tower was input into the model using 1 1 

representative floor plans from CAD, including 

the four transfer floors. Mechanical floors were 

input into the model; however, no occupants 

were placed on these floors. All usable floor 

space had to be filled with nodes measuring 0.5 

m x 0.5 m on which the occupants would travel 

throughout the building. 



On each floor throughout the Tower, three 
stairs were added in their appropriate position 
on the floor and number of risers, depending 
upon the floor. These stairs were added to the 
model with the appropriate diagonal length, 
stair width, door width into and out of the 
stairwell, number of landings, and accurate stair 
riser and tread distances. 

Figure D-3 shows a representation of a stair 
configuration of a 1.1 m (44 in.), 19 riser stair 
(Stairs A or C) in the buildingEXODUS model. 
The solid blue nodes represent the landing 
space and the green patterned nodes represent 
the stair steps. The figure shows a "spiral" 
stair that is 2 nodes wide with the distances in 



0.31 m on 
diagonal of 
stair nodes 




0.6 m 

between 

centers of 

landing nodes 



Y 

0.6 m between 

centers of landing 

nodes 



Figure D-3. 44 in stair configuration in 
buildingEXODUS. 



248 



NIST NCSTAR 1-7, WTC Investigation 



Egress Modeling 



between the centers of the nodes. The occupants can also move diagonally between the outer and inner 
track of the stair, which is not shown in the figure to prevent confusion. Also, the dotted lines in between 
the landing nodes are used to represent that one side of the landing will be located on one floor and the 
other side of the landing will be located on the floor above. Due to the 7.5 in by 9.5 in riser and tread 
configuration (not including the nose extending the tread space), the diagonal space between steps 
measured approximately 12 in. Similar stair configuration was used for the 56 in. stair (Stair B). The 
only difference was the distances from node to node; for instance, instead of a 0.6 m spacing between the 
centers of landing nodes, there was a 0.7 m spacing input. The number of nodes specified along the width 
of the stairwell is the buildingEXODUS's method of simulating stair width, and since not more than one 
person is allowed to occupy the same node, specifying 2 nodes across each stairwell allows for 2-person 
movement abreast in all three stairwells. NIST is aware of the narrowness of the 44 in. stair (Stair A and 
C), however, did not find it appropriate to model only 1 -person abreast inside the stairwell throughout the 
entire building. Therefore, two nodes were used across each stair as an assumption made when using the 
buildingEXODUS model. Stair node connection lengths do not imply areas of the stairs, and therefore 
density of the space, but rather the distance that occupants travel inside the stair. 

For the population, either 19,800 occupants or 25,500 occupants (including visitors) were distributed 
randomly throughout the building space, depending on the scenario. Similar to the phased evacuation 
scenario, the following occupant types were used to describe the WTC tower population: 5 percent Males 
age 17-29, 38 percent Males age 30-50, 21 percent Males age 51-80, 3 percent Females age 17-29, 
22 percent Females age 30-50, and 1 1 percent Females age 51-80. All occupants were able-bodied 
without mobility impairments. 

The minimum evacuation time from the Tower was simulated by choosing an immediate response time 
and the stair pack option, which predicted that occupants will "pack" inside the stairs during evacuation 
(however in the simulation each occupant still tries to leave the step ahead of them empty). Also, it was 
predicted that occupants traveled to the nearest stairwell on each floor. The simulation predicted that 
occupants from a fully-occupied building without visitors, with a response delay of zero, evacuated the 
building in approximately 110 minutes. Also, when visitors were added to the simulation, 
buildingEXODUS predicted an evacuation time of 141 minutes. 

Another group of simulations provided the evacuation time from the Tower if the occupants delayed their 
evacuation by to 10 minutes. All other options chosen for the group of simulations were similar to 
those described for the non delay simulations. buildingEXODUS predicted that occupants from a fully- 
occupied building without visitors, with a response delay of to 10 minutes, evacuated the building in 
approximately 112 minutes. Also, when visitors were added to the simulation, buildingEXODUS 
predicted an evacuation time of 142 minutes. 

EXIT89 

The EXIT89 model was also used to simulate Scenarios 2 and 3, a total building evacuation of a fully- 
occupied tower both with and without the inclusion of visitors. Without visitors, the population totaled 
19,800 people and with visitors, the population totaled 25,500 people. The tower was input into the 
model by describing each of the 1 1 representative floor plans with a series of nodes and arcs. On the 
mechanical floors, only the stair nodes and arcs were associated with these floors, and no occupants were 



NIST NCSTAR 1-7, WTC Investigation 249 



Appendix D 



placed on these floors. All usable floor space, including stairs and transfer corridors, was filled with 
larger-sized nodes ranging from 2 m 2 to 104 m 2 and connected by arcs. 

Similar to buildingEXODUS, three stairs were added in their appropriate position on each floor 
throughout the tower. These stairs were added to the model with the appropriate diagonal length, stair 
width, and door width into and out of the stairwell. 

The stair input for the EXIT89 model was different than that used by the buildingEXODUS model. 
Instead of the stair consisting of two lanes of nodes, a stairwell in EXIT89 was represented by a 
horizontal area (m 2 ), stair width measured where stair nodes meet, and the stair travel distance measured 
along the diagonal, which included the lengths traveled on the landings. Even though the stair input 
method varied between buildingEXODUS and EXIT89, stair distances were verified as being similar in 
both models for the Tower geometry. 

Within EXIT89, the transfer corridors were input as horizontal nodes connecting vertical stair nodes at 
each transfer floor, also part of the geometry input. At each transfer floor, the stairs leading to and 
leading away from the floor were considered completely different stairwells and named accordingly. 
However, since EXIT89 could account for only 10 different stairwells, the smallest transfer at floor 42 
was considered part of a stairwell. 

The number of occupants intentionally placed on each floor within the input file corresponded to the 
occupant numbers calculated per zone from both the 14 m 2 /person and 1 1 m 2 /person densities, depending 
upon the scenario. All occupants were assigned the medium body size and both emergency and normal 
speed options (from multiple simulations as shown by the range of evacuation times in Table D-5). 

The simulation predicted that occupants from a fully-occupied building without visitors, with a response 
delay of zero, evacuated the building in approximately 97 to 117 minutes (emergency to normal speed 
input conditions). Also, when visitors were added to the simulation, EXIT89 predicted an evacuation 
time of 1 14 to 140 minutes. 

When a response delay of 0-10 minutes was introduced to EXIT89, the simulation predicted that 
occupants from a fully-occupied building without visitors evacuated the building in approximately 92 to 
113 minutes. Also, when visitors were added to the simulation, EXIT89 predicted an evacuation time of 
119 to 139 minutes. 



250 NIST NCSTAR 1-7, WTC Investigation 



Egress Modeling 



Table D-5. Total Building Evacuation Time (Simulated) for Various Occupant Loads 



Evacuation Model 


Evacuation 

Initiation Delay 

Input 


Evacuation Time 

(min): 8,800 

occupants 


Evacuation time 

(min): 19,800 

occupants 


Evacuation time 

(min): 25,500 

occupants 


EXODUS 


0- 10 Minute Delay 


55 


112 


142 


No Delay 


52 


110 


141 


EXIT89 


0- 10 Minute Delay 


71-74 


92-113 


119-139 


No Delay 


58-78 


97 69 -117 


114-140 



D.4.3 Scenario Four (September 11th Capacity) Inputs and Results 

The buildingEXODUS and EXIT89 models were used to perform the total building evacuation analysis 
of a WTC tower, Scenario 4, with a population similar to that observed on the morning of 
September 11, 2001. All inputs for the geometry and simulation options in buildingEXODUS and 
EXIT89 were identical to those used in Scenarios 2 and 3. Also, the inputs for the population options 
were identical, except for the number of occupants simulated. For scenario four, a total building 
evacuation was simulated with 8,800 occupants (the capacity estimated for a tower on the morning of 
September 11,2001). 

buildingEXODUS predicted that occupants from a building holding a September 11, 2001, capacity 
evacuated the building in approximately 52 minutes with no delay and 55 minutes with a to 10 minute 
delay time. EXIT89 predicted that occupants from a building holding a September 11, 2001, capacity 
evacuated the building in approximately 58 to 78 minutes (depending up the speed option chosen, 
emergency to normal) with no delay and 71 to 74 minutes with a to 10 minute delay time. 

D.4.4 Scenarios Five and Six (Full-Capacity with Damage) Inputs and Results 

To model how the event would have changed if the buildings were fully occupied, models were run using 
the calculated fully-occupied number of occupants in each tower, 19,800 occupants (without visitors). 
Through running these scenarios, NIST hoped to answer the following questions: 

• What would have been the consequences if WTC 1 had been fully-occupied on September 11, 
2001, the building had been hit by aircraft, and then had collapsed 103 minutes later? 

• What would have been the consequences if WTC 2 had been fully-occupied on September 11, 
2001, the building had been hit by aircraft 16 minutes after WTC 1 was hit, and then had 
collapsed 72 minutes after WTC 1 was hit (56 minutes from the time that WTC 2 was hit)? 

To model these scenarios, NIST ran a series of calibration simulations for each tower in an attempt to 
crudely simulate the occupant conditions inside the towers on September 11, 2001 with the capacity of 



The underlying theory for people movement in EXIT89 is based upon the work of Predtechenskii and Milinskii. They 
observed an inverse relationship between density and speed for three different types of movement: emergency, normal, and 
comfortable. The effect of crowding (density) on overall evacuation speed was greater when no delay time was assumed. A 
delay time when 19,800 occupants were present, according to EXIT89, spaces the occupants out and increases overall 
evacuation efficiency. This was not found to be the case for all occupant loads (25,500). 



NISTNCSTAR 1-7, WTC Investigation 



251 



Appendix D 

8,800 occupants. Once the calibration simulations approximately achieved the goal of simulating the 
September 11, 2001 scenario, the fully-occupied population was added to each tower. From the fully- 
occupied simulations of each tower, possible consequences were predicted, such as the number of 
occupants who would be trapped inside the tower at the time of collapse and the congestion points (the 
Mezzanine escalator) that developed with a larger population. 

For the calibration simulations, there were some difficult aspects of the evacuation that needed to be 
accounted for in the evacuation, such as pre-evacuation activities, occupants changing stairs for various 
reasons and the activities that they perform on the floor during evacuation, occupants resting in the stairs, 
occupants being delayed by helping behaviors, occupants giving way to injured persons (superflow), and 
firefighters counterflow. Although it was recognized that the models cannot accurately simulate all of 
these activities as of yet, an attempt was made to account for time lost to perform such activities in the 
calibration simulations. The buildingEXODUS model was the primary model used for this set of 
scenarios. Due to the way that the stairs are created in the model (lanes of occupants instead of an 
occupied space), it was difficult to model actual firefighters walking throughout the building and achieve 
the same kind of occupant response and effect that were seen on September 11, 2001. It was also difficult 
to simulate the effect of a larger bodied individual on the rest of the occupants in the stair, due to the 
method of simulating stair movement used in the buildingEXODUS model (2-lane stairways). 

The successful calibration simulation used for the WTC involved a method (to be referred to as) the "stop 
and go" method. The stop and go method involved the creation of "doors" inside the stair that would 
open and close throughout the simulation. This method attempted to account for evacuation interruptions 
and the "stop and go" reports from survivors during the evacuation at different points throughout the stair, 
such as leaving/changing stairs (sometimes occupants did this multiple times) for various reasons, resting 
on the stairs, helping, waiting behind larger or disabled occupants, superflow, firefighter counterflow, etc. 
Evacuation delays were distributed uniformly throughout the building, depending upon the Tower, to 
account for pre-evacuation delays and any actions that occupants took when leaving the stairs, such as 
short activities like moving directly to the next stairs, and longer activities such as seeking information, 
making phone calls, looking for staff/coworkers, and looking out windows. In each building, a minimum 
evacuation delay of 6 minutes was assigned to the population because even if the occupant left their floor 
immediately, they were still faced with other challenges/actions taken while outside of the stairs. 
Although the evacuation delay distribution on September 1 1 , 200 1 , was not uniform, this assumption was 
made for the model. 

One must remember that the "stop and go" method was only an estimate of the reported "stop and go" 
activities observed in the WTC towers. This "stop and go" method alleviated a large crowd surrounding 
the escalator, which was a result in the simulations run for Scenarios 2-4, but was not overwhelmingly 
reported by survivors of September 11, 2001. 

D.4.5 WTC 1 Calibration Simulation 

The inputs for the calibration simulation of WTC 1 were the following: 
• 7,200 occupants distributed from floors 90-ground floor 



252 NIST NCSTAR 1-7, WTC Investigation 



Egress Modeling 



"Stop and go doors" located inside the stairwells at floor 76 (early in the evacuation) to simulate 
the locked doors in the transfer, floor 5 (late in the evacuation), and floors 15 and 48 during the 
evacuation 



• 



• 6-30 minute delay uniformly distributed over the entire population 

• Distribution of gender and age used in Scenarios 2-4 

• Stair packing and local potential 

In the calibration simulation, 7,200 occupants were distributed from floor 90 to the ground floor. The 
"stopping" points occurred (door closed) at every 5 minute interval for 2 minutes at floors 15 and 48 
starting after 10 minutes. Then, at floor 76, the "door closed" only three times for 2 minutes each, to 
simulate the locked doors on floor 76 in Stair A at the beginning of the simulation. Lastly, on floor 5, the 
door closed 5 times for 2 minutes each at the ending time of the evacuation. Also, a 6 - 30 minute delay 
was uniformly distributed to the entire population of the building. 

The simulation was judged to be "successful" at predicting a September 11,2001, scenario if the 
following criteria set for the WTC 1 calibration simulation were met: 

• Evacuated more than 90 percent but less than 100 percent of the occupants from the tower before 
WTC 2 collapse (72 min) 

• Alleviated significant crowding at the entrance to the Mezzanine escalator 

For the calibration simulation, when 7,200 occupants in WTC 1 were given a 6-30 minute time delay with 
stoppage points, no significant queue developed at the escalator. Also, 7,200 occupants (under a uniform 
delay) evacuated the building in approximately 71 minutes. This is a good estimate for WTC 1, because 
90 percent of the occupants reported leaving the WTC complex before WTC 2 collapsed. Therefore, this 
geometry and scenario set up was used to simulate a fully-occupied WTC 1, which contained 
16,000 occupants below floor 92 (19,800 occupants minus those trapped above floor 91). Of the 
remaining 1 percent who were trying to evacuate, however, some were located inside the towers at the 
very base of the building (which the model falls a bit short of predicting), and some were located in the 
Concourse area trying to evacuate the complex at this time. Further, this simulation did not account for 
those who waited until WTC 2 collapsed to initiate their evacuation, (e.g., occupants on floor 64, and 
some very high in the building trapped by significant floor damage). 

D.4.6 WTC 1 Full-Capacity with Damage Simulation 

When 16,000 occupants were simulated using the calibrated input assumptions, buildingEXODUS 
calculated that there were still 2,400 occupants left inside WTC 1 as the building began to collapse. 
There were larger queues simulated at the escalator in a full-capacity scenario when compared to the 
September 11, 2001, calibration simulation. It is understood that with a larger population, occupants 
would have encountered even more stoppages in the stairs than occupants did on September 11, 2001; 
however, this is not quantifiable and only the delay times and stoppages used in the calibration simulation 
were used in this scenario. Stoppage points in the full-capacity simulation continued until 10 minutes 
after the collapse of WTC 2. 

NIST NCSTAR 1-7, WTC Investigation 253 



Appendix D 

D.4.7 WTC 2 Calibration Simulation 

The inputs for the calibration simulation of WTC 2 were the following: 

7,400 occupants distributed from floors 107-ground floor (after subtracting elevator users). 

"Stop and go doors" located inside the stairwells at floor 5 (late in the evacuation), and floor 15 
and 48 during the evacuation, beginning after 20 minutes 

"Door" closed inside stairs at floor 78 after 16 minutes 

2-17 minute delay uniformly distributed over population from floors 107-78; 6-30 minute delay 
uniformly distributed over the population from floors 77 - ground 

Distribution of gender and age used in Scenarios 2-4 

Stair packing and local potential 

A total of 7,400 occupants (84 percent of 8,800) were specifically placed throughout each of the three 
building sections, depending upon the percentage of occupants reporting elevator usage within that 
section. Stoppage points were placed low in the building to simulate stopping for fire officials, 
superflow, helping behaviors, etc. on floors 15 and 48 (throughout the evacuation beginning at 
20 minutes) and floor 5 (toward the end of the simulation). The reason that the stoppages began after 
20 minutes is because the fire department and others did not respond to WTC 2 until the building was hit. 

This simulation modeled evacuation from the entire building until WTC 2 was attacked. After 
16 minutes, it was assumed that everyone above the impact floor, 78, would not be able to evacuate past 
the impact point, even though it is known that 1 8 occupants escaped from this area. The reason for 
assigning a difference in delay times to occupants above and below the impact zone was because the time 
delay was used to account for both pre-evacuation activities and activities performed during the 
evacuation, such as waiting on a floor. Since occupants above the impact zone were not able to complete 
their evacuation, a lower delay range was assigned to them. 

The simulation was judged to be "successful" at predicting a September 11, 2001, scenario if the 
following criteria set for the WTC 2 calibration simulation were met: 

• All occupants below the 78th floor evacuated before T2 collapses (72 min) 

• A loss of approximately 600-700 occupants was predicted above 78 when WTC 2 collapsed 

• Alleviated significant crowding at the entrance to the Mezzanine escalator 

The WTC 2 calibration simulation predicted that all occupants evacuate from below floor 78 and that 
645 occupants remained above floor 78 when WTC 2 collapsed. Also, there was no significant queuing 
predicted at the entrance to the Mezzanine escalator. These modeling assumptions were then used to 
simulate the full-capacity tower containing 17,260 occupants, which accounted for occupants who used 
elevators, as described below. 



254 NIST NCSTAR 1-7, WTC Investigation 



Egress Modeling 



Elevators 

For WTC 2, a fully occupied building on September 11, 2001 was assumed to contain 19,800 occupants. 
It was necessary to estimate the number of occupants who would have been able to use the elevators 
within the first 16 minutes on September 11, 2001, when WTC 2 was fully occupied. Calculations were 
performed using the ELVAC model (Klote 1993). From the estimations made by the face-to-face 
interview respondents using the shuttle elevators (NIST 2004), the peak speed of the shuttle elevators was 
approximately 4.8 m/s (close to 78 floors in 60 seconds). The ELVAC model was used to calculate how 
many occupants could have evacuated to the ground floor of WTC 2 in less than 1 6 minutes using 
80 percent of the shuttle elevators available on floors 44 and 78. Only 80 percent of the elevators were 
modeled, under the assumption that some elevators would be out of service. Under the assumptions that 
the peak elevator speed was 4.8 m/s and had an acceleration of 1.5 m/s 2 , 20 percent of the occupants 
(1,337) from the high section of the building (floors 78-110), 25 percent of the occupants (1,661) from the 
middle section of the building (floors 44-77), and 6 percent of the occupants (388) from the lower section 
of the building would have been able to evacuate via elevators before their tower was hit. 

The maximum number of occupants who would have been able to use the shuttle elevators, which had a 
capacity of 50 occupants each, to evacuate the building in under 14 minutes was estimated to be 
3,386 occupants. However, since it was discovered from face-to-face interviews (NIST 2004) that some 
elevators traveled to the ground floor without full capacity, it was assumed that only 75 percent of the 
maximum number of occupants who could use elevators (0.75 x 3,386 = 2,540) evacuated using the 
elevators. Therefore, approximately 17,260 occupants used the stairs in this simulation (19,800 - 2,540 = 
17,260). 

D.4.8 WTC 2 Full-Capacity with Damage Simulation 

When 17,260 occupants were placed within the calibration simulation, buildingEXODUS predicted that 
there were still over 8,000 occupants left inside WTC 2 after 72 minutes. Occupants were distributed 
throughout the building, above and below the floors of impact, according to predicted elevator use, and 
assigned an appropriate response delay. As was found with the WTC 1 full-capacity simulation, a larger 
amount of queuing was predicted by the buildingEXODUS model, especially at the entrance to the 
Mezzanine escalator. 

Table D-6 shows the results for the calibration and fully-occupied scenarios with building damage for 
both WTC 1 and WTC 2. 



NISTNCSTAR 1-7, WTC Investigation 255 



Appendix D 



Table D-6. WTC 1 and WTC 2 full capacity evacuation simulation results. 



Simulation 
Title 


Geometry; 
Population 


Simulation 


# occupants evacuated 
before tower collapses 


WTC 1 
Calibration 9/11 


Floors 90-ground; 
7,200 occupants 


Response time = 6-30 min, stoppage 
points throughout the stairs to simulate 
stopping of occupants for approx. 2 
minute intervals 


7,200 evacuate from below 
impact region within 71 
minutes 


WTC 1 fully- 
occupied 
building on 9/1 1 


Floor 90-ground; 
1 6,000 occupants 


Response time = 6-30 min, stoppage 
points throughout the stairs to simulate 
stopping of occupants for approx. 2 
minute intervals 


13,600 occupants evacuated 
before WTC 1 collapsed; 
85 % evacuated from below 
impact zone 


WTC 2 
Calibration 9/11 


Floors 107-ground; 
7,400 occupants 
when removing 
occupants using 
elevators 


Response time 2-17 minutes above 78; 
6-30 minutes all other places; stoppage 
points throughout the stairs that begin 
after 20 minutes, "door" closes inside 
stair at 78 after 16 minutes 


6,755 occupants evacuated 
before WTC 2 collapsed; 
645 remained above floor 
78 


WTC 2 fully- 
occupied 
building on 9/1 1 


Floors 107-ground; 
1 7,260 occupants 
when removing 
occupants using 
elevators 


Response time = 6-30 min, stoppage 
points throughout the stairs to simulate 
stopping of occupants for approx. 2 
minute intervals 


8,883 occupants evacuate 
before WTC 2 collapsed; 
3,900 remained above floor 
78; 58% evacuated the 
WTC 2 



Differences were found between the results of the full-capacity with damage simulations for WTC 1 and 
WTC 2. The reasons for the differences in results are the following: 

• In WTC 1, all "stop and go" points within each stairwell were ceased 10 minutes after the 
collapse of WTC 2. This allowed "free flow" from 82 minutes to 103 minutes in WTC 1. In 
WTC 2, the firefighters had no prior warning to stop counterflow measures and begin to evacuate. 
Therefore, the "stop and go" points continued until the collapse of WTC 2. 

• Occupants in WTC 1 had a total of 103 minutes before their building collapsed, whereas 
occupants in WTC 2 had only 72 minutes. 

• In WTC 2, with a much higher population, there were a greater number of congestion points 
throughout the building, in addition to the "stop and go" points added to the building. 
Additionally, some occupants higher in the building were given a longer evacuation time delay to 
account for activities performed during evacuation, causing them to take longer than the allotted 
72 minutes to reach the ground floor. These congestion points were located at the transfer floors. 

• Only 58 percent of the occupants in the fully-occupied building evacuated WTC 2, via the use of 
stairs or elevators. 2,540 were removed from the evacuation because they used elevators and 
8,883 occupants evacuated via the stairs from WTC 2, for a total of 1 1,423 out of 19,800 
occupants. 



256 



NIST NCSTAR 1-7, WTC Investigation 



Egress Modeling 



D.5 



CONCLUSIONS AND FINDINGS 



As shown in Table D-7, four different scenarios were run using three evacuation models, with and 
without a 0-10 minute delay time randomly distributed among the occupants. The first column of results 
shows the times for phased evacuation (with and without a delay time of 0-10 minutes) and the last three 
columns show full evacuations of occupants with different building populations. On average, the phased 
evacuation time (with initiation delays) for three floors to evacuate to three floors below the fire floor was 
1 1 minutes. However, for a fully-occupied building (19,800 occupants), evacuation times were calculated 
as 8 to 10 times larger for a total evacuation when compared to phased evacuation for this building. 
These scenarios assume a to 10 minute delay time and do not account for other types of behavior that 
may delay evacuation and increase total evacuation time. 





Table D-7 


. Summary ol 


1 egress simulat 


on results. 




Evacuation 
Model 


Evacuation 

Initiation Delay 

Input 


Phased 

Evacuation: 

600 occupants 


Evacuation 

Time (min): 

8,800 occupants 


Evacuation time 

(min): 19,800 

occupants 


Evacuation time 

(min): 25,500 

occupants 


EXODUS 


0-10 Minute Delay 


11 


55 


112 


142 


No Delay 


4 


52 


110 


141 


EXIT89 


0-10 Minute Delay 


11.5 


71-74 


92-113 


119-139 


No Delay 


3.5 


58-78 


97 70 -117 


114-140 


Simulex 


0-10 Minute Delay 


11.5 




No Delay 


4 



Some mid-rise buildings that plan for phased evacuation can completely evacuate their building, at full 
capacity, in less than 20 minutes. However, a building such as a World Trade Center tower produces full 
evacuation times of over 100 minutes, which is a large difference in comparison to the phased evacuation 
time of 1 1 minutes. With buildingEXODUS, adding a 0-10 minute delay did not significantly affect the 
evacuation times from the building, adding only two additional minutes. The movement inside and from 
the base of the stairwell was the limiting factor and controlled the evacuation time. When the starting 
population roughly doubled, the evacuation time roughly doubled. As this simulation did not account for 
behavioral aspects of the evacuation (such as leaving the stairs and resting), the full impact may be 
underestimated somewhat by these evacuation model simulations. 

Scenarios 2-4 estimated the total evacuation time for a fully-occupied building on September 11, 2001. 
Figure D-4 shows that the evacuation with 25,585 occupants took approximately 2.5 times longer to 
evacuate than the 8,800 occupant population. Therefore, to estimate the total evacuation time from a 
fully-occupied building under the conditions observed on September 11, 2001, multiply by 2.5 the 
observed time required to evacuate WTC 1 on September 1 1 (approximately 100 minutes). This estimate 
is subject to the assumption that upon adding additional occupants, the emergency evacuation time scales 



1 The underlying theory for people movement in EXIT89 is based upon the work of Predtechenskii and Milinskii. They 
observed an inverse relationship between density and speed for three different types of movement: emergency, normal, and 
comfortable. The effect of crowding (density) on overall evacuation speed was greater when no delay time was assumed. A 
delay time when 19,800 occupants were present, according to EXIT89, spaces the occupants out and increases overall 
evacuation efficiency. This was not found to be the case for all occupant loads (25,500). 



NISTNCSTAR 1-7, WTC Investigation 



257 



Appendix D 



linearly, as was observed with the modeled evacuation time. A fully occupied building with visitors, but 
without elevator usage, would have required approximately four hours (2.5 x 100 minutes « 4 hours ) to 
evacuate the building. Using the same method, if WTC 1 was occupied with approximately 20,000 
occupants (without visitors), the evacuation would have required roughly twice as long to evacuate than 
the 8,800 occupant simulation. Therefore, a fully occupied building without visitors and without the use 
of elevators would have required over three hours (2 x 1 00 minutes « 3 hours ) to evacuate the building. 
Note that in 2005, PANYNJ estimated that the maximum population of WTC 1 or WTC 2 would not 
likely have exceeded 20,000. 71 



Full Building Evacuation of a WTC Tower 



160 



140 



c 


120 


E 




a> 


100 


b 




H 




c 


80 


o 








(8 

3 


60 


o 




cB 




> 
HI 


40 



20 



25,585 - full capacity 
with visitors 



19,800 - full capacity 



8,800 - 9/11 capacity 



5000 10000 15000 20000 

Number of occupants in Tower 



25000 



30000 



Figure D-4. Full building evacuation simulation results with differing building 

populations. 

Scenarios 5-6 were modeled to estimate the consequences of a fully-occupied building on 
September 11, 2001 by estimating the number of lives lost at each building collapse. As the results show, 
more people would have lost their lives, particularly in WTC 2 due to its earlier collapse. Table D-8 
shows the simulated number of occupants (by starting location) who successfully evacuated WTC 1 and 
WTC 2 for a fully-occupied building subject to the conditions observed on September 11, 2001. 

• WTC 1-69 percent of all occupants (13,600 occupants out of 19,800) would have successfully 
evacuated the Tower, which breaks down to percent (0 out of 3,800) from above the impact 
zone and 85 percent (13,600 out of 16,000) from below the impact zone in 103 minutes. 



Bhol, Saroj. PANYNJ (September 21, 2005). Email from S. Bhol to S. Sunder in response to NIST question. 



258 



NIST NCSTAR 1-7, WTC Investigation 



Egress Modeling 



• WTC 2-58 percent of all occupants (8,883 by stairs and 2,540 by elevators) would have 

successfully evacuated the tower with the use of stairs and elevators in 72 minutes, which breaks 
down to 44 percent from above the impact zone and 75 percent from below the impact zone in 72 
minutes. 



Table D-8. Simulated evacuation results for fully-occupied WTC 1 and WTC 2 on 

September 11, 2001. 



Building 


Total 

Number of 

Occupants 

at t = 0.0 


Potential 
Number 

of 
Evacuees 


Occupants 
Remaining 
in Building 
at Collapse 


Occupants 
Trapped 

Above 

Floors of 

Impact 


Percentage of Occupants Who 

Successfully Evacuated Relative to 

Where They Started 


Total 


Below 
Impact 


Above 
Impact 


WTC 1 


19,800 


16,000 


6,200 


3,800 


69% 


85% 


0% 


WTC 2 


19,800 


17,260 


8,377 


3,900 


58% 


75% 


44% 


Total 


39,600 


33,260 


14,577 


7,700 


63% 







NISTNCSTAR 1-7, WTC Investigation 



259 



Appendix D 



This page intentionally left blank. 



260 NIST NCSTAR 1-7, WTC Investigation